1 00:00:02,800 --> 00:00:07,440 Speaker 1: Zoe Media. Hello one, Welcome to this week's Better Offline Monologue. 2 00:00:07,480 --> 00:00:18,400 Speaker 1: I'm your host, led Zitron. Now this week's monologue is 3 00:00:18,440 --> 00:00:20,920 Speaker 1: about something I also just through up in my newsletter, 4 00:00:21,160 --> 00:00:24,080 Speaker 1: A new term I came up with called analoslav. Now 5 00:00:24,120 --> 00:00:28,080 Speaker 1: analoslov is when somebody writes a long, specious, but authoritative 6 00:00:28,120 --> 00:00:31,080 Speaker 1: sounding piece of writing with few facts are actual statements 7 00:00:31,240 --> 00:00:34,440 Speaker 1: with the intention of it being read as thorough analysis, 8 00:00:34,640 --> 00:00:38,360 Speaker 1: research and analysis. Indeed, now this week, alleged research group 9 00:00:38,400 --> 00:00:41,320 Speaker 1: Cittrini Research, not to be confused with Andrew left Citron 10 00:00:41,360 --> 00:00:45,160 Speaker 1: Research all well me put out a truly awful piece 11 00:00:45,200 --> 00:00:48,640 Speaker 1: of writing called the twenty twenty eight Global Intelligence Crisis, 12 00:00:48,800 --> 00:00:52,360 Speaker 1: which is this slop filled scare fiction written and framed 13 00:00:52,360 --> 00:00:55,360 Speaker 1: with the authority of deeply founded analysis, so much so 14 00:00:55,480 --> 00:00:57,840 Speaker 1: that it caused the global cell off and stocks. And 15 00:00:57,880 --> 00:01:00,480 Speaker 1: I must be clear if you took this seriously, if 16 00:01:00,480 --> 00:01:02,680 Speaker 1: you listen to this on I don't know you had 17 00:01:02,680 --> 00:01:05,640 Speaker 1: it read out by LM notebook, or you read it 18 00:01:05,680 --> 00:01:07,760 Speaker 1: and you kind of furwed your brow and like this, 19 00:01:08,280 --> 00:01:11,440 Speaker 1: this is really provocative. You're a mark. You are being conned, 20 00:01:11,440 --> 00:01:13,479 Speaker 1: You're in the process of being conned. If you haven't 21 00:01:13,480 --> 00:01:17,000 Speaker 1: already been now this piece. If you haven't read it, 22 00:01:17,080 --> 00:01:18,880 Speaker 1: please do go and read. It will be linked in 23 00:01:18,920 --> 00:01:21,680 Speaker 1: the notes my annotated version. Those of you in YouTube 24 00:01:21,680 --> 00:01:23,720 Speaker 1: won't get the link. Just go to my Blue Sky 25 00:01:23,760 --> 00:01:26,760 Speaker 1: I swear to god it's there. It spends seven thousand 26 00:01:26,880 --> 00:01:29,640 Speaker 1: words telling the dire tale of what would happen if 27 00:01:29,680 --> 00:01:34,240 Speaker 1: AI made an indeterminately large amount of white collar workers redundant. 28 00:01:34,760 --> 00:01:39,639 Speaker 1: Now you read this seven thousand boring, fucking word long piece, 29 00:01:40,000 --> 00:01:43,880 Speaker 1: it isn't clear what AI exactly does, who makes the AI, 30 00:01:44,040 --> 00:01:47,120 Speaker 1: how the AI works, really anything about it, just that 31 00:01:47,200 --> 00:01:51,520 Speaker 1: it replaces people and then bad stuff happens. Citrini insists 32 00:01:51,640 --> 00:01:54,600 Speaker 1: that this isn't bare porn or AI doomer fan fiction, 33 00:01:54,880 --> 00:01:58,080 Speaker 1: but that's exactly what it is. Mediocre analyslot framed in 34 00:01:58,120 --> 00:02:01,000 Speaker 1: the trappings of analysis, sold on a subset with researching 35 00:02:01,040 --> 00:02:05,680 Speaker 1: the title specifically written and successfully spooking and ingratiating anyone 36 00:02:05,800 --> 00:02:08,240 Speaker 1: involved in the financial markets who wants to be scared 37 00:02:08,240 --> 00:02:10,880 Speaker 1: of this stuff. Its goal is to convince you that 38 00:02:10,960 --> 00:02:14,720 Speaker 1: AI non specifically is scary, that your current stocks are bad, 39 00:02:14,760 --> 00:02:17,320 Speaker 1: and that AI stocks unclear which ones those are, by 40 00:02:17,360 --> 00:02:19,960 Speaker 1: the way, are the future also find out more for 41 00:02:20,040 --> 00:02:22,280 Speaker 1: nine hundred and ninety nine dollars a year. I tried 42 00:02:22,400 --> 00:02:25,000 Speaker 1: seventy bucks, and my shit's way better than this fuck 43 00:02:25,040 --> 00:02:31,079 Speaker 1: when Now look, look, okay, let me give you an example. 44 00:02:31,200 --> 00:02:33,120 Speaker 1: I'm going to do a voice. It should have been 45 00:02:33,120 --> 00:02:35,560 Speaker 1: clearer all along that a single GPU cluster in North 46 00:02:35,600 --> 00:02:39,119 Speaker 1: Dakota generating the output previously attributed to ten thousand white 47 00:02:39,120 --> 00:02:42,600 Speaker 1: collar workers in midtown Manhattan is more economic pandemic than 48 00:02:42,720 --> 00:02:47,360 Speaker 1: am economic panacea. Now, he writes economic correctly, I just 49 00:02:47,480 --> 00:02:50,160 Speaker 1: fucked up saying it. Now. The goal of a paragraph 50 00:02:50,200 --> 00:02:52,280 Speaker 1: like this is for you to say, Wow, that's what 51 00:02:52,360 --> 00:02:56,040 Speaker 1: GPUs are doing. Now, they're doing that today. It isn't. 52 00:02:56,120 --> 00:02:59,160 Speaker 1: Of course, the majority of CEOs report little or no 53 00:02:59,240 --> 00:03:01,760 Speaker 1: return on investment from AI, with a study of six 54 00:03:01,840 --> 00:03:05,359 Speaker 1: thousand CEOs across the US, UK, Germany and Australia finding 55 00:03:05,400 --> 00:03:08,519 Speaker 1: that more than eighty percent detected no discernible impact from 56 00:03:08,520 --> 00:03:12,720 Speaker 1: AI on either employment or productivity. Nevertheless, you read GPU 57 00:03:12,800 --> 00:03:14,919 Speaker 1: and North Dakota and you think, wow, that's the place 58 00:03:14,960 --> 00:03:18,000 Speaker 1: I know, and I know that GPU's power AI. Now 59 00:03:18,080 --> 00:03:20,520 Speaker 1: I'm a curious little critter, so I know a GPU 60 00:03:20,560 --> 00:03:23,760 Speaker 1: cluster in North Dakota Corweeb's won with applied Digital that 61 00:03:23,800 --> 00:03:26,720 Speaker 1: has debts so severe that loses both companies money even 62 00:03:26,760 --> 00:03:29,000 Speaker 1: if they have the capacity rented out twenty four to seven. 63 00:03:29,400 --> 00:03:31,480 Speaker 1: But let's not let facts get in the way of 64 00:03:31,520 --> 00:03:34,360 Speaker 1: a poorly written story. Now, I don't need to go 65 00:03:34,440 --> 00:03:36,480 Speaker 1: line by line on this. I'm not going to because 66 00:03:36,480 --> 00:03:39,560 Speaker 1: I'll end up saying a legally actionable threat of some kind. 67 00:03:39,760 --> 00:03:41,240 Speaker 1: But I need you to know that most of this 68 00:03:41,320 --> 00:03:44,920 Speaker 1: piece's arguments come down to magical thinking and utterly empty pros. 69 00:03:45,280 --> 00:03:48,119 Speaker 1: For example, how do you think that AI takes over 70 00:03:48,160 --> 00:03:53,760 Speaker 1: the entire economy? And I quote AI capabilities improved, Companies 71 00:03:53,760 --> 00:03:57,640 Speaker 1: needed fewer workers, white collar layoffs increased, displaced workers spent 72 00:03:57,720 --> 00:04:00,680 Speaker 1: less margin pressure pushed firms to invest more in AI. 73 00:04:00,920 --> 00:04:06,400 Speaker 1: AI capabilities improved. That's right, they just get better. Now, 74 00:04:06,360 --> 00:04:09,440 Speaker 1: I need to discuss anything happening today. Even AI twenty 75 00:04:09,480 --> 00:04:11,880 Speaker 1: twenty seven had the balls to start making stuff up 76 00:04:11,880 --> 00:04:15,520 Speaker 1: about open brain or some such bullshit. This piece literally 77 00:04:15,560 --> 00:04:20,080 Speaker 1: just says stuff, including this particularly egregious lin I'm about 78 00:04:20,120 --> 00:04:23,159 Speaker 1: to read to you. In late twenty twenty five, agentic 79 00:04:23,240 --> 00:04:26,400 Speaker 1: coding tools to cold step function jump in capability. A 80 00:04:26,400 --> 00:04:29,160 Speaker 1: competent developer working with cloud code or Codex could now 81 00:04:29,200 --> 00:04:32,240 Speaker 1: replicate the core functionality of a mid market SaaS product 82 00:04:32,240 --> 00:04:36,159 Speaker 1: in weeks, not perfectly or with every single edge case handled, 83 00:04:36,160 --> 00:04:39,080 Speaker 1: but well enough that the CIO reviewing a five hundred 84 00:04:39,080 --> 00:04:41,880 Speaker 1: thousand annual renewal started asking the question, what if we 85 00:04:42,040 --> 00:04:47,719 Speaker 1: just built this ourselves now, quick thing? What annual contract? 86 00:04:47,760 --> 00:04:50,520 Speaker 1: There are tons of SaaS contracts, though far longer than that. 87 00:04:50,680 --> 00:04:52,520 Speaker 1: But this is also a complete and not a lie, 88 00:04:52,560 --> 00:04:55,119 Speaker 1: a ball faced lie. This is not something that claud 89 00:04:55,200 --> 00:04:57,520 Speaker 1: code can do. The fact that we have major media 90 00:04:57,560 --> 00:05:01,320 Speaker 1: outlets quoting this piece suggests that those responsible for explaining 91 00:05:01,320 --> 00:05:03,800 Speaker 1: how things work don't actually bother to do any of 92 00:05:03,920 --> 00:05:06,240 Speaker 1: the work to find out. And it's both a disgrace 93 00:05:06,240 --> 00:05:08,640 Speaker 1: and an embarrassment for the tech and business media that 94 00:05:08,680 --> 00:05:11,480 Speaker 1: these lies continue to be peddled. As it went over 95 00:05:11,520 --> 00:05:14,800 Speaker 1: in last week's monologue, the entire AI replacing software story 96 00:05:14,880 --> 00:05:17,000 Speaker 1: is a con It's clear that the markets and an 97 00:05:16,960 --> 00:05:20,000 Speaker 1: alarming amount of people in the media simply do not 98 00:05:20,200 --> 00:05:23,320 Speaker 1: know what they are talking about, or are intentionally avoiding 99 00:05:23,400 --> 00:05:27,800 Speaker 1: thinking about it. The AI replaces software story is literally 100 00:05:27,880 --> 00:05:31,279 Speaker 1: Anthropic has released a product and now the resulting industry 101 00:05:31,960 --> 00:05:34,720 Speaker 1: is selling off, such as when it launched a cybersecurity 102 00:05:34,760 --> 00:05:37,240 Speaker 1: tool that could check for vulnerabilities, a product that's existed 103 00:05:37,240 --> 00:05:39,960 Speaker 1: in some form for nearly a decade, and this caused 104 00:05:39,960 --> 00:05:43,080 Speaker 1: to sell off in cybersecurity stocks like CrowdStrike, you know, 105 00:05:43,160 --> 00:05:45,120 Speaker 1: the one that had a faulty bit of code cause 106 00:05:45,160 --> 00:05:48,880 Speaker 1: a global cybersecurity incident that lost the fortune five hundred 107 00:05:48,960 --> 00:05:51,720 Speaker 1: billions and resulted in Delta Airlines having to cancel over 108 00:05:51,760 --> 00:05:55,479 Speaker 1: twelve hundred flights over a period of several days. There 109 00:05:55,560 --> 00:05:59,320 Speaker 1: is no rational basis for anything about this selloff other 110 00:05:59,360 --> 00:06:01,919 Speaker 1: than that off financial media and markets do not appear 111 00:06:01,960 --> 00:06:05,160 Speaker 1: to understand the very basic things about the stuff they 112 00:06:05,200 --> 00:06:08,520 Speaker 1: invest in or talk about. Software may seem complex, but 113 00:06:08,800 --> 00:06:12,840 Speaker 1: especially in these cases, it's really quite simple. Investors are 114 00:06:12,839 --> 00:06:15,440 Speaker 1: conflating an AI model can spit out code with an 115 00:06:15,480 --> 00:06:18,760 Speaker 1: AI model can create an entire experience of what we 116 00:06:18,800 --> 00:06:21,599 Speaker 1: know as software, or is close enough to that we 117 00:06:21,680 --> 00:06:26,239 Speaker 1: have to start freaking out, which it isn't it obviously isn't. 118 00:06:26,839 --> 00:06:29,640 Speaker 1: Anyone who build software knows this. Anyone peddling this is 119 00:06:29,680 --> 00:06:33,760 Speaker 1: just wrong, and it's this replet vibe coding shit. I 120 00:06:33,800 --> 00:06:35,680 Speaker 1: swear to God. People in the media use it and 121 00:06:35,720 --> 00:06:38,240 Speaker 1: they just go, Wow, this is the future now or 122 00:06:38,320 --> 00:06:40,160 Speaker 1: software will be built like this as they fail to 123 00:06:40,200 --> 00:06:43,800 Speaker 1: build any calm down at come down the software? Why not? 124 00:06:43,920 --> 00:06:48,640 Speaker 1: You know? I think we really need to think deeply 125 00:06:49,160 --> 00:06:51,320 Speaker 1: about how. For the second time in a month, the 126 00:06:51,360 --> 00:06:53,880 Speaker 1: markets in the media have had a miniature shit fit 127 00:06:54,120 --> 00:06:57,560 Speaker 1: based on blogs that tell lies using fan fiction. As 128 00:06:57,560 --> 00:07:00,839 Speaker 1: I covered in my annotations of Matt Schumer's all Something 129 00:07:00,880 --> 00:07:03,440 Speaker 1: Big Is Happening piece, the people that are meant to 130 00:07:03,440 --> 00:07:05,920 Speaker 1: tell the general public what's happening in the world appear 131 00:07:05,960 --> 00:07:08,799 Speaker 1: to be falling for ghost stories that confirm their biases 132 00:07:08,880 --> 00:07:11,640 Speaker 1: or investment strategies, even if said stories are full of 133 00:07:11,680 --> 00:07:16,760 Speaker 1: half truths and outright lies. I'm despairing a little when 134 00:07:16,760 --> 00:07:19,160 Speaker 1: I see Matt Schumer on CNN or hear from the 135 00:07:19,160 --> 00:07:21,680 Speaker 1: head of a p firm about Citrini research. I begin 136 00:07:21,760 --> 00:07:24,760 Speaker 1: to wonder whether everybody got where they were not through 137 00:07:24,800 --> 00:07:28,080 Speaker 1: any actual work or knowledge, but by making the right noises. 138 00:07:29,080 --> 00:07:31,679 Speaker 1: We're in a grifter economy, and the people that should 139 00:07:31,680 --> 00:07:34,720 Speaker 1: be stopping the grifters are asleep at the fucking wheel.