1 00:00:02,800 --> 00:00:08,320 Speaker 1: Zone Media. Hello and welcome to this week's Better Offline Monologue. 2 00:00:08,360 --> 00:00:18,880 Speaker 1: I'm your host ed Zitron. A lot of people have emailed, 3 00:00:18,920 --> 00:00:21,400 Speaker 1: messaged and plugged me that the natural end of the 4 00:00:21,440 --> 00:00:24,520 Speaker 1: AI bubble is some kind of government bailouts, something resembling 5 00:00:24,560 --> 00:00:28,760 Speaker 1: the TARP program that bailed out financial institutions overleveraged during 6 00:00:28,800 --> 00:00:31,240 Speaker 1: the Great Financial Crisis. And I want to be clear 7 00:00:31,280 --> 00:00:34,240 Speaker 1: about something. This is different, and a bailout is very 8 00:00:34,360 --> 00:00:38,000 Speaker 1: very very unlikely. As a reminder, TOP stood for Troubled 9 00:00:38,040 --> 00:00:41,720 Speaker 1: Asset Relief Program. It existed, unsurprisingly, to buy the distressed 10 00:00:41,720 --> 00:00:44,800 Speaker 1: assets that banks had invested in during the Great Financial Crisis, 11 00:00:44,960 --> 00:00:47,879 Speaker 1: primarily the ridiculous bets made on the housing market made 12 00:00:47,960 --> 00:00:50,040 Speaker 1: up of people who had got mortgages they could not 13 00:00:50,120 --> 00:00:54,280 Speaker 1: pay for. Now, these bailouts existed to stabilize banks that 14 00:00:54,320 --> 00:00:56,920 Speaker 1: were going out of business, along with consumers' money being 15 00:00:56,920 --> 00:00:59,080 Speaker 1: set on fire if they did so, as well as 16 00:00:59,120 --> 00:01:01,960 Speaker 1: auto manufacturer like GM and Chrysler that the contagion had 17 00:01:02,000 --> 00:01:03,800 Speaker 1: spread to. And while I'm not going to go into 18 00:01:03,840 --> 00:01:06,360 Speaker 1: massive amounts of detail, haven't got all day, I can 19 00:01:06,400 --> 00:01:09,920 Speaker 1: make it pretty simple. These bailouts existed to stabilize systemically 20 00:01:10,000 --> 00:01:12,800 Speaker 1: necessary companies that had they died, would have tanked the 21 00:01:12,920 --> 00:01:15,399 Speaker 1: US economy and markets at the same time and also 22 00:01:15,520 --> 00:01:19,759 Speaker 1: lost people's money. Unlike the Great Financial Crisis, very few 23 00:01:19,800 --> 00:01:23,000 Speaker 1: of the major players in AI are either systemically necessary 24 00:01:23,080 --> 00:01:24,800 Speaker 1: or in danger of dying as the result of the 25 00:01:24,840 --> 00:01:29,600 Speaker 1: AI bubble bursting. Furthermore, there really isn't anything too bailout. 26 00:01:30,040 --> 00:01:32,640 Speaker 1: So let's start with Nvidia, the largest company on the 27 00:01:32,640 --> 00:01:36,080 Speaker 1: stock market. While GPU sales make up woods of eighty 28 00:01:36,080 --> 00:01:38,520 Speaker 1: eight percent of its revenue and Video will not die 29 00:01:38,560 --> 00:01:40,640 Speaker 1: if the AI bubble bursts, though it will see a 30 00:01:40,680 --> 00:01:43,960 Speaker 1: massive contraction as a result of its biggest business line collapsing. 31 00:01:44,600 --> 00:01:46,760 Speaker 1: It's also important to know that in Nvidia's massive stock 32 00:01:46,800 --> 00:01:50,040 Speaker 1: price is a direct result of its credible growth. It 33 00:01:50,160 --> 00:01:53,240 Speaker 1: isn't enough for in Vidia to continue selling lots of GPUs. 34 00:01:53,440 --> 00:01:55,840 Speaker 1: It must continue to increase revenues by at least sixty 35 00:01:55,880 --> 00:01:59,240 Speaker 1: percent year of a year, every single quarter, which means 36 00:01:59,280 --> 00:02:01,440 Speaker 1: that for in Video to continue being the most beloved 37 00:02:01,480 --> 00:02:04,280 Speaker 1: boy and stock in the in history of the stock market, 38 00:02:04,480 --> 00:02:07,360 Speaker 1: it needs to make ninety one billion dollars in revenue 39 00:02:07,480 --> 00:02:09,720 Speaker 1: in a year's time, and that's just in one quarter. 40 00:02:09,800 --> 00:02:13,880 Speaker 1: By the way, this isn't something a bailout can fix. 41 00:02:14,000 --> 00:02:17,200 Speaker 1: Even in the most crony capitalist fantasy you can think of. 42 00:02:17,800 --> 00:02:20,560 Speaker 1: A bailout would have to continually sink tens of billions 43 00:02:20,600 --> 00:02:24,040 Speaker 1: of dollars directly into Invidia, buying GPUs and increasing amounts 44 00:02:24,240 --> 00:02:28,120 Speaker 1: every three months in perpetuity to continue its valuation. And 45 00:02:28,120 --> 00:02:30,680 Speaker 1: while some of you may say, booh, troub will do 46 00:02:30,840 --> 00:02:35,160 Speaker 1: themon you'd live in a boring, dumerous fancy based on nothing. 47 00:02:35,200 --> 00:02:38,680 Speaker 1: And I'm sick of hearing about it. Good lord, listen 48 00:02:38,760 --> 00:02:41,880 Speaker 1: to me. Top was a seven hundred billion dollar program 49 00:02:41,880 --> 00:02:44,520 Speaker 1: that only ended up investing around four hundred and forty 50 00:02:44,520 --> 00:02:48,119 Speaker 1: three point five billion dollars or so, or put another way, 51 00:02:48,280 --> 00:02:50,480 Speaker 1: less than than the five hundred billion dollars of orders 52 00:02:50,480 --> 00:02:53,880 Speaker 1: that Invidia CEO Jensen Je claims are on Invidia's books 53 00:02:53,880 --> 00:02:56,560 Speaker 1: through the end of twenty twenty six. Are we meant 54 00:02:56,560 --> 00:02:59,280 Speaker 1: to do this forever? Is that the plan? Are you 55 00:02:59,320 --> 00:03:00,760 Speaker 1: saying that Trump is just going to go. 56 00:03:00,720 --> 00:03:02,880 Speaker 2: There, We're going to buy gps or everyone, We're going 57 00:03:02,960 --> 00:03:05,440 Speaker 2: to bury them every year every year? Obey so many 58 00:03:05,440 --> 00:03:07,360 Speaker 2: of the most of you, No, it's not gonna happen. 59 00:03:07,600 --> 00:03:09,080 Speaker 2: Though I guess he has kind of done that with 60 00:03:09,120 --> 00:03:11,960 Speaker 2: the Department of Energy. You've probably heard of this deal. 61 00:03:12,000 --> 00:03:14,040 Speaker 2: By the way, it's one hundred thousand GPUs. That isn't 62 00:03:14,040 --> 00:03:14,720 Speaker 2: going to do shit. 63 00:03:15,280 --> 00:03:17,799 Speaker 1: Jensen Wong needs all your money and eats it now 64 00:03:17,840 --> 00:03:24,200 Speaker 1: every quarter forever. Okay, Okay, it's not going to happen now, Okay. 65 00:03:24,240 --> 00:03:28,120 Speaker 1: Another annoying talking point you've probably heard is that open 66 00:03:28,160 --> 00:03:30,399 Speaker 1: ai is somehow too big to fail, when in fact 67 00:03:30,440 --> 00:03:33,320 Speaker 1: it's too small to pull apart into enough parts to eat. 68 00:03:34,040 --> 00:03:37,800 Speaker 1: Open Ai has promised one point four trillion dollars in 69 00:03:37,840 --> 00:03:41,280 Speaker 1: compute deals, convincing Oracle and others to build massive data 70 00:03:41,320 --> 00:03:43,840 Speaker 1: centers in its honor. Yet the reality is that this 71 00:03:43,880 --> 00:03:46,560 Speaker 1: company isn't even going to hit its thirteen billion dollars 72 00:03:46,640 --> 00:03:49,920 Speaker 1: in projected revenue this year, and we'll burn billions of 73 00:03:49,920 --> 00:03:52,240 Speaker 1: dollars in inference loan and doesn't really have a path 74 00:03:52,240 --> 00:03:53,080 Speaker 1: to profitability. 75 00:03:53,480 --> 00:03:54,000 Speaker 3: Two big to. 76 00:03:53,960 --> 00:03:56,560 Speaker 1: Fail means that something may kill our economy or markets 77 00:03:56,560 --> 00:03:58,640 Speaker 1: if it dies, And as loud and annoying as Clammy 78 00:03:58,680 --> 00:04:02,000 Speaker 1: sam Utman might be, open AI's death wouldn't kill either 79 00:04:02,040 --> 00:04:06,320 Speaker 1: of those. Let's break it down. Open Ai dying would 80 00:04:06,320 --> 00:04:08,920 Speaker 1: be symbolically lethal to the large language model era and 81 00:04:08,920 --> 00:04:11,880 Speaker 1: create an immediate and permanent chill in AI investments. Don't 82 00:04:11,880 --> 00:04:14,280 Speaker 1: get me wrong. Its death would also likely be an 83 00:04:14,280 --> 00:04:16,960 Speaker 1: existential threat to Crusoe, the company building its data center 84 00:04:16,960 --> 00:04:19,640 Speaker 1: in Abilene, Texas for Oracle. I spoke to someone on 85 00:04:19,680 --> 00:04:21,600 Speaker 1: the ground there and they say, if the money runs out, 86 00:04:21,640 --> 00:04:23,279 Speaker 1: everyone just stops and goes home. 87 00:04:23,400 --> 00:04:24,240 Speaker 3: That'll be fine. 88 00:04:24,520 --> 00:04:27,080 Speaker 1: I don't know the extent of Cruser's revenues, however, but 89 00:04:27,120 --> 00:04:28,640 Speaker 1: I do know that Oracle is on the hook for 90 00:04:28,680 --> 00:04:32,200 Speaker 1: a billion dollars a year for Abilene for fifteen goddamn years. 91 00:04:32,720 --> 00:04:35,160 Speaker 1: An Oracle has also taken on massive amounts of debt 92 00:04:35,200 --> 00:04:38,440 Speaker 1: to build Abilene and other data centers in Shackleford, Texas, 93 00:04:38,440 --> 00:04:41,000 Speaker 1: and I think somewhere in Wisconsin is in New Mexico 94 00:04:41,160 --> 00:04:44,400 Speaker 1: one now and started this with an eighteen billion dollar 95 00:04:44,440 --> 00:04:47,040 Speaker 1: bond sale and in the works eighty eight billion dollar 96 00:04:47,080 --> 00:04:49,880 Speaker 1: debt package to build data centers for Open AI. Well, 97 00:04:49,920 --> 00:04:52,680 Speaker 1: Oracle does run the risk of default and credit default swaps, 98 00:04:52,680 --> 00:04:55,159 Speaker 1: which is people betting again well betting that they might 99 00:04:55,279 --> 00:04:59,400 Speaker 1: default on their obligations. That's going up. The risk is there, 100 00:05:00,080 --> 00:05:02,520 Speaker 1: but there's still a real company that makes real things. 101 00:05:03,040 --> 00:05:06,560 Speaker 1: Microsoft owns open Ai IP and research, which would mean 102 00:05:06,600 --> 00:05:09,279 Speaker 1: that Microsoft pretty much absorbed those eight one hundred million 103 00:05:09,320 --> 00:05:12,480 Speaker 1: weekly active users directly into Copilot. And that's if it 104 00:05:12,600 --> 00:05:15,600 Speaker 1: even chose to retain them and just shut down chat 105 00:05:15,640 --> 00:05:19,080 Speaker 1: GBT directed it to a very flimsy version of Copilot 106 00:05:19,080 --> 00:05:20,480 Speaker 1: and didn't let people do half the. 107 00:05:20,440 --> 00:05:22,080 Speaker 3: Shit they did with them. 108 00:05:22,760 --> 00:05:25,960 Speaker 1: Bailing out open Ai would not do anything, in part 109 00:05:26,000 --> 00:05:29,400 Speaker 1: because open ai burns billions of dollars and doesn't really 110 00:05:29,800 --> 00:05:33,360 Speaker 1: do much in the large economy itself. It's revenues, which 111 00:05:33,400 --> 00:05:36,440 Speaker 1: I estimate to be most four point five billion dollars 112 00:05:36,440 --> 00:05:37,520 Speaker 1: through the end of September. 113 00:05:37,880 --> 00:05:39,600 Speaker 3: By the way, if you're listening to this and you. 114 00:05:39,600 --> 00:05:42,640 Speaker 1: Think that my revenue numbers for open Ai, my inference 115 00:05:42,720 --> 00:05:47,000 Speaker 1: numbers were delayed somehow, it's a cruel accounting. Stop fucking 116 00:05:47,040 --> 00:05:49,200 Speaker 1: with me. I know what I'm doing. Anyway, those revenues 117 00:05:49,240 --> 00:05:51,839 Speaker 1: are inconsequential, and the eight point six to seven billion 118 00:05:51,880 --> 00:05:54,080 Speaker 1: dollars it spent on inference in that period is while 119 00:05:54,080 --> 00:05:57,240 Speaker 1: a large chunk of Vizio revenue from Microsoft only really 120 00:05:57,279 --> 00:06:00,080 Speaker 1: accounting for three point eight eight percent or so, the 121 00:06:00,080 --> 00:06:02,479 Speaker 1: two hundred and twenty three billion dollars Microsoft has made 122 00:06:02,480 --> 00:06:05,800 Speaker 1: in that period. Losing open AI would hurt but hardly 123 00:06:05,839 --> 00:06:08,680 Speaker 1: threaten the health of anyone in Redmond. The same ghost 124 00:06:08,720 --> 00:06:12,640 Speaker 1: for Anthropic or any other generative AI company. These companies 125 00:06:12,640 --> 00:06:16,080 Speaker 1: aren't really spending that much, nor are they fueling much 126 00:06:16,440 --> 00:06:20,080 Speaker 1: of any actual real economic activity. There's only around sixty 127 00:06:20,120 --> 00:06:22,839 Speaker 1: billion dollars of actual AI revenue in twenty twenty five. 128 00:06:23,040 --> 00:06:25,600 Speaker 1: In a tech industry where Amazon made one hundred and 129 00:06:25,640 --> 00:06:28,440 Speaker 1: eighty billion dollars and Google one hundred and two billion 130 00:06:28,480 --> 00:06:32,320 Speaker 1: dollars in their last quarters. The disappearance of llm's writ 131 00:06:32,480 --> 00:06:35,120 Speaker 1: large would have a miniscule effect on actual real dollars 132 00:06:35,240 --> 00:06:39,520 Speaker 1: entering the economy, other than investments in data centers, which 133 00:06:39,560 --> 00:06:43,320 Speaker 1: is where I think there's, if any, some kind of bailout. 134 00:06:44,680 --> 00:06:48,240 Speaker 1: As I've mentioned recently, there's been about fifty billion dollars 135 00:06:48,760 --> 00:06:51,080 Speaker 1: or more even every quarter for the last three quarters 136 00:06:51,080 --> 00:06:53,200 Speaker 1: going into building data centers, and I think that many 137 00:06:53,240 --> 00:06:57,120 Speaker 1: of these projects end up defaulting. There's a chance a 138 00:06:57,320 --> 00:07:00,360 Speaker 1: chance that these are somehow bailed out, if an only 139 00:07:00,400 --> 00:07:03,599 Speaker 1: if they provide an actual existential risk to a major 140 00:07:03,640 --> 00:07:06,680 Speaker 1: asset firm, A Blackstone or areas of what have you. 141 00:07:07,040 --> 00:07:08,400 Speaker 1: I don't know if Areas is in them and just 142 00:07:08,480 --> 00:07:11,840 Speaker 1: naming them at this point. Nevertheless, there is one that 143 00:07:11,920 --> 00:07:15,360 Speaker 1: I the one company I could see if any bailout happens. 144 00:07:15,360 --> 00:07:18,320 Speaker 1: And I'm really hesitant to say I think there's any chance. 145 00:07:19,200 --> 00:07:23,040 Speaker 1: Oracle Oracle I think is the one chance. Oracle is 146 00:07:23,040 --> 00:07:25,400 Speaker 1: currently massively leveraged and going into debt to the point 147 00:07:25,440 --> 00:07:27,400 Speaker 1: it's eating into its free cash flow, and it's a 148 00:07:27,440 --> 00:07:29,800 Speaker 1: close ally of the Trump administration. I could see them 149 00:07:29,800 --> 00:07:33,720 Speaker 1: getting a lifeline to escape these ridiculous loans. And while 150 00:07:33,760 --> 00:07:36,040 Speaker 1: you may say there's some sort of grand plan or 151 00:07:36,040 --> 00:07:38,480 Speaker 1: the administration will agree to pay these companies to keep 152 00:07:38,480 --> 00:07:40,960 Speaker 1: them alive, I don't see it happening. Is doing so 153 00:07:41,120 --> 00:07:43,840 Speaker 1: won't support the economy, the markets really do anything other 154 00:07:43,880 --> 00:07:49,080 Speaker 1: than prolongedly inevitable, It's not going to happen, and I 155 00:07:49,080 --> 00:07:51,000 Speaker 1: I don't know. I'm I want to phrase this in 156 00:07:51,040 --> 00:07:52,520 Speaker 1: a nice way because I don't want it seem like 157 00:07:52,560 --> 00:07:55,000 Speaker 1: an attack. Every time you say something like this, you're 158 00:07:55,000 --> 00:07:57,440 Speaker 1: doing marketing for the AI industry. 159 00:07:57,760 --> 00:07:59,360 Speaker 3: Dumerism is marketing. 160 00:07:59,480 --> 00:08:03,040 Speaker 1: At this point, I'm so scared of what they'll they'll create, 161 00:08:03,240 --> 00:08:05,840 Speaker 1: they are gonna get bail out open Aye is too 162 00:08:05,840 --> 00:08:06,520 Speaker 1: big to fail. 163 00:08:06,760 --> 00:08:09,680 Speaker 3: All these are extrapolations of your fear. I get it. 164 00:08:09,840 --> 00:08:13,160 Speaker 1: They're also marketing. You are helping them. You are helping 165 00:08:13,200 --> 00:08:17,200 Speaker 1: them market their services by making them seem systemically relevant 166 00:08:17,200 --> 00:08:19,920 Speaker 1: when they are not, by making them seem like they're 167 00:08:19,920 --> 00:08:24,360 Speaker 1: innovative when they're not. Honestly, I need you to just 168 00:08:24,400 --> 00:08:26,680 Speaker 1: live in reality and look at what's actually happening, which 169 00:08:26,720 --> 00:08:30,240 Speaker 1: is sweet fuck all. In many cases, every one of 170 00:08:30,280 --> 00:08:32,440 Speaker 1: these wanky studies you hear come out so on, it's 171 00:08:32,480 --> 00:08:35,079 Speaker 1: like ten percent of jobs can already be automated by AI. 172 00:08:35,360 --> 00:08:37,959 Speaker 1: CNBC should be fucking ashamed of themselves for running a 173 00:08:37,960 --> 00:08:38,600 Speaker 1: story about it. 174 00:08:38,600 --> 00:08:39,280 Speaker 3: It's a PhD. 175 00:08:39,640 --> 00:08:42,680 Speaker 1: Congrats to the Better Offline Reddit user resource. It's a 176 00:08:42,679 --> 00:08:47,800 Speaker 1: PhD paper. And also it's referring to random skills rather 177 00:08:47,840 --> 00:08:50,760 Speaker 1: than jobs, and skills and jobs are two different fucking things. 178 00:08:51,240 --> 00:08:56,280 Speaker 1: I'm so goddamn tired of this era, good Lord, But nevertheless, 179 00:08:56,559 --> 00:08:58,520 Speaker 1: the next two weeks are going to be really interesting. 180 00:08:58,559 --> 00:09:00,360 Speaker 1: I'm working on a three or four part or about 181 00:09:00,360 --> 00:09:02,319 Speaker 1: my worries around in video, and I'm going to tell 182 00:09:02,320 --> 00:09:05,400 Speaker 1: you all about in Video, and how not similar they 183 00:09:05,440 --> 00:09:08,559 Speaker 1: are to Enron and Video is nothing like Enron or 184 00:09:08,640 --> 00:09:12,240 Speaker 1: Loosen or Welcome or WinStar or all these other companies. 185 00:09:12,280 --> 00:09:14,640 Speaker 3: And we're going to go into as well. And tomorrow I've. 186 00:09:14,480 --> 00:09:18,560 Speaker 1: Got a fun interview with Nathan Grayson of Aftermath Your Love. 187 00:09:19,240 --> 00:09:23,440 Speaker 1: I might next week take a knee and just run 188 00:09:23,520 --> 00:09:26,319 Speaker 1: the better offline in studio thing I just did with 189 00:09:26,840 --> 00:09:29,319 Speaker 1: Lisa from CNN, or I might do the three or 190 00:09:29,320 --> 00:09:31,360 Speaker 1: four bar. It really comes down to how I'm feeling. 191 00:09:31,440 --> 00:09:34,400 Speaker 1: It's been a rough few weeks. Health wise. I'm doing fine. 192 00:09:34,400 --> 00:09:36,720 Speaker 1: I've just been sick and I've just been quite stressed. 193 00:09:37,120 --> 00:09:39,720 Speaker 1: I must again say thank you for all the lovely 194 00:09:39,760 --> 00:09:42,120 Speaker 1: emails and messages you've been sending. 195 00:09:42,160 --> 00:09:43,600 Speaker 3: You're all so lovely. I love you all. 196 00:09:44,320 --> 00:09:47,240 Speaker 1: Couldn't be lucky with the listeners I've got. Tell your 197 00:09:47,240 --> 00:09:49,640 Speaker 1: friends about the show, tell them about the newsletter, tell 198 00:09:49,679 --> 00:09:53,160 Speaker 1: them about me, tell them about my strange proclivities, but 199 00:09:53,240 --> 00:09:55,640 Speaker 1: also keep listening. I really do appreciate all of you 200 00:09:56,080 --> 00:09:58,679 Speaker 1: enjoyed tomorrow's episode and whatever I put out next week, 201 00:09:58,720 --> 00:10:01,000 Speaker 1: and there will be a monologue regard Godless, unless I 202 00:10:01,080 --> 00:10:02,760 Speaker 1: do the three or four par in which case, that's 203 00:10:02,760 --> 00:10:04,680 Speaker 1: going to be a monologue for the week. And yes, 204 00:10:04,800 --> 00:10:06,920 Speaker 1: by the way, I realize me just talking is always 205 00:10:06,960 --> 00:10:07,520 Speaker 1: a monologue. 206 00:10:07,559 --> 00:10:10,079 Speaker 3: I get it. You're so smart. You are though, really 207 00:10:10,200 --> 00:10:11,040 Speaker 3: thank you for listening.