1 00:00:02,480 --> 00:00:07,440 Speaker 1: All Zone Media. Hello, and welcome to Better Offline Live 2 00:00:07,560 --> 00:00:09,760 Speaker 1: and direct, coming at you from the middle of nowhere 3 00:00:10,080 --> 00:00:12,840 Speaker 1: the center of everywhere, and I'm your host ed zetron. 4 00:00:21,440 --> 00:00:26,760 Speaker 1: What else it feels like forever since I talked directly 5 00:00:26,840 --> 00:00:29,440 Speaker 1: to you, my dear listeners. So I'll get right to it. 6 00:00:29,680 --> 00:00:32,200 Speaker 1: And as a reminder, I'll be including links to everything 7 00:00:32,240 --> 00:00:35,040 Speaker 1: I'm talking about in the episode notes. And if you're 8 00:00:35,080 --> 00:00:37,840 Speaker 1: just catching up, all you really need to know here 9 00:00:38,120 --> 00:00:41,400 Speaker 1: is that open ai makes large language models like GPT 10 00:00:41,560 --> 00:00:43,920 Speaker 1: and of course the product chat GPT and throp it 11 00:00:44,000 --> 00:00:47,680 Speaker 1: makes claude a similar product, and that generative AI products 12 00:00:47,680 --> 00:00:50,080 Speaker 1: have yet to really prove a use case that justifies 13 00:00:50,080 --> 00:00:55,720 Speaker 1: them losing money on every single transaction. Let's begin. August second, 14 00:00:55,800 --> 00:00:58,560 Speaker 1: twenty twenty four was Black Friday for the AI boom, 15 00:00:58,640 --> 00:01:00,680 Speaker 1: as a week of rough earnings from big tech led 16 00:01:00,720 --> 00:01:04,600 Speaker 1: to what felt like the entire media industry asking one question, 17 00:01:05,319 --> 00:01:08,840 Speaker 1: is the AI bubble popping? And that's the question I'm 18 00:01:08,840 --> 00:01:10,880 Speaker 1: going to try and answer for you today and in 19 00:01:10,920 --> 00:01:14,640 Speaker 1: the next episode two, the Guardians sought to answer why 20 00:01:14,640 --> 00:01:16,919 Speaker 1: the big seven tech companies had been hit with AI 21 00:01:16,959 --> 00:01:20,680 Speaker 1: boomed outs. CNN asked has the AI bubble burst, and 22 00:01:20,720 --> 00:01:23,720 Speaker 1: The Atlantic suggested several months too late that the generative 23 00:01:23,720 --> 00:01:27,960 Speaker 1: AI revolution may indeed be a bubble. The Financial Times 24 00:01:27,959 --> 00:01:31,959 Speaker 1: reported hedge fund Elliott Management told investors that Nvidia was 25 00:01:32,040 --> 00:01:35,000 Speaker 1: a bubble, and Bloomberg reported that big tech had failed 26 00:01:35,040 --> 00:01:39,039 Speaker 1: to convince Wall Street that AI was paying off. It's 27 00:01:39,080 --> 00:01:41,640 Speaker 1: all part of a growing trend where people are suddenly 28 00:01:41,680 --> 00:01:45,720 Speaker 1: realizing what I've been saying for months, that this unprofitable, 29 00:01:45,840 --> 00:01:49,560 Speaker 1: energy hungry technology that creates mediocre outputs is not actually 30 00:01:49,600 --> 00:01:53,800 Speaker 1: the future. It's just another cloud do hickey. I had 31 00:01:53,800 --> 00:01:55,840 Speaker 1: not a great way of putting it, I realized, but 32 00:01:56,760 --> 00:01:59,480 Speaker 1: this is my nice way of saying. Generative AI isn't 33 00:01:59,520 --> 00:02:03,440 Speaker 1: completely useless, but the large language models they kind of 34 00:02:03,480 --> 00:02:06,480 Speaker 1: are based on the costs. These articles were for the 35 00:02:06,480 --> 00:02:09,120 Speaker 1: most part, talking about the sudden and violent declines and 36 00:02:09,160 --> 00:02:11,960 Speaker 1: the share prices of companies like Microsoft, Amazon, and Google, 37 00:02:12,200 --> 00:02:15,440 Speaker 1: which for the most part haven't really recovered and as 38 00:02:15,480 --> 00:02:18,840 Speaker 1: of about August tenth, are down significantly over the past month. 39 00:02:19,720 --> 00:02:22,680 Speaker 1: Although these articles didn't say anything particularly new or reveal 40 00:02:22,720 --> 00:02:25,760 Speaker 1: any individual missteps or scandals that might have prompted the slide, 41 00:02:26,040 --> 00:02:28,520 Speaker 1: they hinted to a broader awareness among Wall Street that 42 00:02:28,560 --> 00:02:31,600 Speaker 1: the AI ambitions of these companies and by which I 43 00:02:31,639 --> 00:02:35,600 Speaker 1: really mean generative AI, will require these massive upfront investments 44 00:02:36,120 --> 00:02:38,960 Speaker 1: and the payoff might not n actually be there, and 45 00:02:39,040 --> 00:02:41,440 Speaker 1: at least if it is, it's not going to be 46 00:02:41,480 --> 00:02:44,919 Speaker 1: there for a while. We're talking years, if not a decade, 47 00:02:44,960 --> 00:02:47,600 Speaker 1: and once the narrative gets settled, is very, very very 48 00:02:47,639 --> 00:02:51,680 Speaker 1: hard to move in. While a stock market isn't always rational, 49 00:02:51,800 --> 00:02:54,560 Speaker 1: check out Tesla for more. The one day sell off 50 00:02:54,600 --> 00:02:57,200 Speaker 1: of these stocks and the comments from analysts and industry 51 00:02:57,200 --> 00:02:59,880 Speaker 1: figures that followed, suggest that Wall Street is growing in 52 00:03:00,000 --> 00:03:02,680 Speaker 1: increasingly uncomfortable with the vast amounts of money required to 53 00:03:02,680 --> 00:03:06,760 Speaker 1: build and grow generative AI into whatever the fuck that's 54 00:03:06,800 --> 00:03:10,600 Speaker 1: meant to be. These companies, while still making over ten 55 00:03:10,639 --> 00:03:13,760 Speaker 1: billion dollars in profit, of course, referring to Microsoft, Google, Amazon, 56 00:03:13,800 --> 00:03:16,440 Speaker 1: and the like in the last quarter alone, have also 57 00:03:16,560 --> 00:03:19,320 Speaker 1: spent an absolute shit ton of money on infrastructure to 58 00:03:19,360 --> 00:03:24,840 Speaker 1: capture the so called demand for cloud services from generative AI. However, 59 00:03:25,840 --> 00:03:28,280 Speaker 1: one little problem, none of them seem to actually be 60 00:03:28,400 --> 00:03:32,320 Speaker 1: making that much money from the thing they're investing in. 61 00:03:32,320 --> 00:03:36,120 Speaker 1: In the last fiscal year, Microsoft's capex the capital expenditures 62 00:03:36,680 --> 00:03:39,200 Speaker 1: was about fifty five point seven billion dollars, which is 63 00:03:39,280 --> 00:03:41,960 Speaker 1: up seventy five percent year over year, with more than 64 00:03:42,000 --> 00:03:45,240 Speaker 1: one third nineteen billion dollars spent in the last quarter 65 00:03:45,360 --> 00:03:48,880 Speaker 1: and in June thirtieth, twenty twenty four. This is reportedly 66 00:03:48,880 --> 00:03:52,360 Speaker 1: split fifty to fifty between infrastructure and technology, which suggests 67 00:03:52,360 --> 00:03:55,360 Speaker 1: an aggressive data center build out, with Chief financial officer 68 00:03:55,400 --> 00:03:59,280 Speaker 1: for Microsoft Amyhood saying that the company expects capital expenditures 69 00:03:59,280 --> 00:04:02,280 Speaker 1: to increase on a sequential basis, giving cloud and the 70 00:04:02,400 --> 00:04:05,480 Speaker 1: and I quote AI demand that, as I've repeatedly said, 71 00:04:05,600 --> 00:04:09,840 Speaker 1: is not really there. And what exactly does increasing on 72 00:04:09,880 --> 00:04:14,760 Speaker 1: a sequential basis mean in dollar terms? I have absolutely 73 00:04:14,800 --> 00:04:17,520 Speaker 1: no idea. It's vague and perhaps vague enough to rattle 74 00:04:17,520 --> 00:04:20,320 Speaker 1: the markets, particularly when Microsoft will be starting from an 75 00:04:20,360 --> 00:04:24,640 Speaker 1: already quite high valuation. We're still, Hood added, and I 76 00:04:24,720 --> 00:04:29,400 Speaker 1: quote Microsoft's earnings calles that AI related spend represented nearly 77 00:04:29,640 --> 00:04:32,760 Speaker 1: all of our total capital expenditures, with roughly half for 78 00:04:32,839 --> 00:04:35,880 Speaker 1: infrastructure needs that will support monetization over the next fifteen 79 00:04:35,920 --> 00:04:40,240 Speaker 1: years and beyond ugh. In essence, Microsoft spent nineteen billion 80 00:04:40,279 --> 00:04:42,760 Speaker 1: dollars in the last quarter on cloud and AI expenses 81 00:04:42,880 --> 00:04:44,800 Speaker 1: and has made it clear that it's not done spending 82 00:04:44,839 --> 00:04:48,160 Speaker 1: more money than it's ever spent before on a technology 83 00:04:48,200 --> 00:04:50,760 Speaker 1: that neither makes Microsoft nor the people paying them that 84 00:04:50,920 --> 00:04:55,479 Speaker 1: much money. It's very stupid and for context, Microsoft made 85 00:04:55,520 --> 00:04:58,760 Speaker 1: twenty two point zero four billion dollars in profits in 86 00:04:58,839 --> 00:05:02,520 Speaker 1: Q two twenty twenty five. Is this really worth sinking 87 00:05:02,520 --> 00:05:05,880 Speaker 1: an entire quarters worth of profits into? Let me give 88 00:05:05,880 --> 00:05:08,400 Speaker 1: it another way to look at it. Microsoft's net profit 89 00:05:08,440 --> 00:05:10,679 Speaker 1: margin has dropped from thirty nine point four four percent 90 00:05:10,720 --> 00:05:13,000 Speaker 1: in Q three twenty twenty three to thirty four point 91 00:05:13,240 --> 00:05:15,680 Speaker 1: four percent in Q two twenty twenty four, meaning that 92 00:05:15,680 --> 00:05:18,800 Speaker 1: it's taking home less money than it usually would because 93 00:05:18,800 --> 00:05:21,359 Speaker 1: they're really investing in this thing that only loses money. 94 00:05:21,400 --> 00:05:24,080 Speaker 1: It's very good, it's very good that this is happening. 95 00:05:24,400 --> 00:05:27,640 Speaker 1: This is what you love to see. The other cloud 96 00:05:27,640 --> 00:05:32,040 Speaker 1: providers aren't really doing much better. Google's capital expenditures are 97 00:05:32,120 --> 00:05:35,280 Speaker 1: forecast to be fifty billion dollars in twenty twenty four, 98 00:05:35,440 --> 00:05:37,799 Speaker 1: and it's spent eleven billion dollars in Q four twenty 99 00:05:37,800 --> 00:05:41,040 Speaker 1: twenty three, driven by and I quote mostly by technical 100 00:05:41,040 --> 00:05:45,640 Speaker 1: infrastructure meaning servers and data centers, and twelve billion dollars 101 00:05:45,920 --> 00:05:47,920 Speaker 1: was what they spent in Q one twenty twenty four. 102 00:05:48,640 --> 00:05:51,080 Speaker 1: The reason I'm not breaking things out much with Google 103 00:05:51,160 --> 00:05:55,000 Speaker 1: is because it hasn't been extremely guarded about its AI expenses, 104 00:05:55,480 --> 00:05:57,960 Speaker 1: probably because they're really high and not making any money 105 00:05:58,000 --> 00:06:01,400 Speaker 1: from it. Amazon's similarly guarded, with its capex last year 106 00:06:01,440 --> 00:06:04,760 Speaker 1: hanging somewhere around forty eight point four billion dollars. It's 107 00:06:04,800 --> 00:06:07,360 Speaker 1: spent thirty point five billion dollars so far in twenty 108 00:06:07,360 --> 00:06:10,960 Speaker 1: twenty four, an absolutely ridiculous amount considering its profit for 109 00:06:11,040 --> 00:06:13,040 Speaker 1: Q two twenty twenty four was thirteen point four to 110 00:06:13,040 --> 00:06:16,880 Speaker 1: eight billion. They're just sinking their profits into these things, 111 00:06:17,360 --> 00:06:21,520 Speaker 1: and there's really no sign that anything changes. Every hyperscaler 112 00:06:21,560 --> 00:06:24,000 Speaker 1: has said that they intend to keep spending all of 113 00:06:24,040 --> 00:06:26,719 Speaker 1: this money on AI, and I haven't even mentioned companies 114 00:06:26,760 --> 00:06:29,240 Speaker 1: like Oracle, which expects to spend ten billion dollars on 115 00:06:29,240 --> 00:06:32,200 Speaker 1: cloud infrastructure this year, with much of that new capacity 116 00:06:32,200 --> 00:06:35,520 Speaker 1: going to support Microsoft and Google Cloud. I don't know 117 00:06:35,600 --> 00:06:38,200 Speaker 1: what happens if they don't need it anymore. Anyone think 118 00:06:38,240 --> 00:06:43,560 Speaker 1: about anyway anyway, But no Black Friday. It was a 119 00:06:43,560 --> 00:06:46,159 Speaker 1: collective realization of the scale and the cost of AI 120 00:06:46,520 --> 00:06:48,760 Speaker 1: and the first signs that the markets are starting to 121 00:06:48,760 --> 00:06:52,360 Speaker 1: ask those annoying little questions about whether it's actually worth it. 122 00:06:53,240 --> 00:06:57,240 Speaker 1: Yet the real chaos, and one that comparatively speaking slipped 123 00:06:57,279 --> 00:06:59,719 Speaker 1: under the radar, came in the form of one of 124 00:06:59,760 --> 00:07:03,279 Speaker 1: my pale horses. Last month, I put out an episode 125 00:07:03,320 --> 00:07:05,679 Speaker 1: called top Culture where I suggested that the first signs 126 00:07:05,680 --> 00:07:08,120 Speaker 1: of the AI bubbles collapse would be in the failure 127 00:07:08,120 --> 00:07:10,880 Speaker 1: of a major AI company, though one not operating at 128 00:07:10,880 --> 00:07:14,480 Speaker 1: the scale of say Open Ai, and I specifically picked 129 00:07:14,520 --> 00:07:16,960 Speaker 1: one out character dot Ai, which raised one hundred and 130 00:07:17,000 --> 00:07:19,800 Speaker 1: fifty million dollars in funding, and the information had already 131 00:07:19,880 --> 00:07:23,640 Speaker 1: hinted might sell itself to one of the big tech companies. Now. 132 00:07:23,640 --> 00:07:25,680 Speaker 1: The reason I picked them is that their app which 133 00:07:25,680 --> 00:07:28,960 Speaker 1: allows you to talk to AI chatbots of Elon Musk 134 00:07:29,000 --> 00:07:34,200 Speaker 1: and Staragojo from jiu Jitsu Kaizen, the anime manga which 135 00:07:34,240 --> 00:07:37,800 Speaker 1: we all know I'm tired of waiting for. That's not 136 00:07:37,880 --> 00:07:40,440 Speaker 1: a business, by the way, and they never really had 137 00:07:40,440 --> 00:07:43,400 Speaker 1: a sustainable business model or really a meaningful product. And 138 00:07:43,440 --> 00:07:45,400 Speaker 1: there one hundred and fifty million dollar raise was the 139 00:07:45,440 --> 00:07:48,600 Speaker 1: first big stupid capital raise in the generative AI boom. 140 00:07:48,720 --> 00:07:51,360 Speaker 1: Early twenty twenty three, when nobody else could raise at all. 141 00:07:52,600 --> 00:07:55,720 Speaker 1: In my opinion, their continued existence, other than being a 142 00:07:55,800 --> 00:07:59,600 Speaker 1: disgusting insult to company building and startups everywhere, was proof 143 00:07:59,600 --> 00:08:02,800 Speaker 1: that the bubble existed, and their death or whatever form 144 00:08:02,920 --> 00:08:05,240 Speaker 1: this really is, is a sign that the tolerance for 145 00:08:05,320 --> 00:08:09,120 Speaker 1: bullshit is leaving the market, as my blunt force for 146 00:08:09,160 --> 00:08:12,360 Speaker 1: shadowing suggests. On Friday August second, and yes, most of 147 00:08:12,400 --> 00:08:15,360 Speaker 1: this stuff happened. On that Friday, Google said that it 148 00:08:15,360 --> 00:08:19,120 Speaker 1: would license Character AI's technology and hire the company's leaders, 149 00:08:19,160 --> 00:08:22,360 Speaker 1: Noam Shazia and Daniel Defritas, along with their research team 150 00:08:22,400 --> 00:08:25,640 Speaker 1: of thirty people, to work at Google's DeepMind AI division. 151 00:08:26,320 --> 00:08:28,440 Speaker 1: The fates of the other one hundred and forty employees 152 00:08:28,480 --> 00:08:32,160 Speaker 1: remain uncertain. I assume they're alive, but I imagine some 153 00:08:32,240 --> 00:08:34,960 Speaker 1: will remain on as staff as the character dot ai 154 00:08:35,000 --> 00:08:39,160 Speaker 1: app remains operational. It's a strange, sad end for a 155 00:08:39,160 --> 00:08:42,000 Speaker 1: company that never really had any business existing, and in 156 00:08:42,040 --> 00:08:44,679 Speaker 1: many ways feels like a con happening in broad daylight. 157 00:08:45,520 --> 00:08:48,680 Speaker 1: Shazir and Defritas are both former Google employees, having left 158 00:08:48,679 --> 00:08:51,800 Speaker 1: the company in twenty twenty one to create Character dot ai. 159 00:08:52,200 --> 00:08:55,640 Speaker 1: It's unclear whether given their previous employment, they'll be required 160 00:08:55,640 --> 00:08:58,080 Speaker 1: to wear the new GLO propeller hat that you get 161 00:08:58,320 --> 00:09:01,120 Speaker 1: if you're new to Google. I actually don't know if 162 00:09:01,160 --> 00:09:03,400 Speaker 1: they still do that. Email me is easy at better 163 00:09:03,480 --> 00:09:05,760 Speaker 1: offline dot com. That's the letter E than the letter 164 00:09:05,840 --> 00:09:09,360 Speaker 1: Z or Z for my British and Canadian fans. If 165 00:09:09,400 --> 00:09:11,959 Speaker 1: you know about this propeller or indeed can get me one, 166 00:09:12,000 --> 00:09:14,920 Speaker 1: I would very much like one anyway. While this is 167 00:09:14,960 --> 00:09:18,319 Speaker 1: being framed as a typical licensing employee poaching deal where 168 00:09:18,600 --> 00:09:22,400 Speaker 1: nothing acquisition adjacent has happened, it is actually an acquisition. 169 00:09:22,840 --> 00:09:26,520 Speaker 1: Google's paying two point five billion dollars to investors. Employee 170 00:09:26,520 --> 00:09:29,200 Speaker 1: stock options will vest, and that means when you get 171 00:09:29,200 --> 00:09:31,120 Speaker 1: stock in a company, it usually takes time for you 172 00:09:31,240 --> 00:09:33,440 Speaker 1: to actually earn it, so that you can't just take 173 00:09:33,480 --> 00:09:36,160 Speaker 1: it and leave it immediately, as I would, And that 174 00:09:36,240 --> 00:09:39,080 Speaker 1: will keep vesting until July twenty twenty six at the 175 00:09:39,120 --> 00:09:41,520 Speaker 1: acquisition rate of eighty eight dollars a share paid for 176 00:09:41,679 --> 00:09:43,800 Speaker 1: by the money from the licensing deal. That is not 177 00:09:43,880 --> 00:09:47,120 Speaker 1: an acquisition from Google. But after that point your shit 178 00:09:47,200 --> 00:09:49,800 Speaker 1: out of luck. It's no longer guaranteed. This is a 179 00:09:50,120 --> 00:09:53,959 Speaker 1: terrible situation. All around for everyone other than the earliest investors. 180 00:09:53,960 --> 00:09:57,600 Speaker 1: It's another great situation where the bad guys win. In 181 00:09:57,720 --> 00:10:00,360 Speaker 1: essence anyone who has options that aren't fully there. By 182 00:10:00,440 --> 00:10:02,880 Speaker 1: July twenty twenty six, maybe shit out of luck and 183 00:10:02,960 --> 00:10:06,480 Speaker 1: character AI's original funding valued them a billion dollars, making 184 00:10:06,480 --> 00:10:09,200 Speaker 1: this a situation where and Resent Horowitz aw a one 185 00:10:09,240 --> 00:10:11,480 Speaker 1: point five x return on their investment in a company 186 00:10:11,480 --> 00:10:15,319 Speaker 1: that never really did anything. Employees kind of got screwed. 187 00:10:15,360 --> 00:10:19,400 Speaker 1: Maybe they didn't, It's unclear, But what is clear is 188 00:10:19,400 --> 00:10:22,840 Speaker 1: that as of now, character ai is effectively dead. The 189 00:10:22,880 --> 00:10:25,640 Speaker 1: company will shift from using their own models their literal 190 00:10:25,720 --> 00:10:28,960 Speaker 1: only thing that they did, to publicly available open source 191 00:10:29,000 --> 00:10:32,600 Speaker 1: large language models like metas Lama, and the original engineering 192 00:10:32,679 --> 00:10:35,719 Speaker 1: team is effectively gone. The company will likely shamble a 193 00:10:35,760 --> 00:10:38,160 Speaker 1: long life lessly until it curls up in the corner 194 00:10:38,200 --> 00:10:40,439 Speaker 1: like an old cat, Except an old cat has more 195 00:10:40,480 --> 00:10:44,640 Speaker 1: purpose and meaning in the universe than the bullshit chatbot company. 196 00:10:45,800 --> 00:10:49,440 Speaker 1: So why does this matter? On one level, the death 197 00:10:49,480 --> 00:10:52,440 Speaker 1: of character dot ai is an indicator of the unsustainability 198 00:10:52,440 --> 00:10:55,800 Speaker 1: of many of these generative AI applications and companies. Even 199 00:10:55,840 --> 00:10:58,160 Speaker 1: with one hundred and fifty million dollars in funding, which 200 00:10:58,200 --> 00:11:00,880 Speaker 1: is a decent amount of money for a normal company. 201 00:11:01,440 --> 00:11:04,240 Speaker 1: Character likely couldn't keep the lights on for very long, 202 00:11:04,600 --> 00:11:06,440 Speaker 1: which kind of made it necessary for them to be 203 00:11:06,520 --> 00:11:10,240 Speaker 1: absorbed or acquired before they shut down and embarrassed everyone involved. 204 00:11:11,280 --> 00:11:15,400 Speaker 1: It's also just like a Charlatan's Olympics. It's this company sucked. 205 00:11:15,440 --> 00:11:19,000 Speaker 1: It never made money, it didn't really do anything that different, 206 00:11:19,960 --> 00:11:23,000 Speaker 1: but they just get bailed out. These fucking start up people. 207 00:11:23,000 --> 00:11:25,439 Speaker 1: They always go on about the meritocracy. They're always going 208 00:11:25,480 --> 00:11:28,559 Speaker 1: on about, Oh, it's about working hard, It's not about 209 00:11:28,559 --> 00:11:32,960 Speaker 1: your circumstances. These jack holes, these fucking idiots. They got 210 00:11:32,960 --> 00:11:35,640 Speaker 1: to sell their bullshit company to Google, a company they 211 00:11:35,720 --> 00:11:38,680 Speaker 1: left the found it. It's just so annoying. And this 212 00:11:38,800 --> 00:11:42,560 Speaker 1: agreement also strongly resembles an earlier one between Inflection AI 213 00:11:42,880 --> 00:11:46,760 Speaker 1: and Microsoft, where Microsoft bought Inflections technology and team without 214 00:11:47,000 --> 00:11:50,520 Speaker 1: actually acquiring its equity, so the actual stock in the company. 215 00:11:51,080 --> 00:11:54,480 Speaker 1: Founded by DeepMind co founder mister Fassuliman and LinkedIn co 216 00:11:54,600 --> 00:11:57,480 Speaker 1: founder Reid Hoffman in twenty twenty two, Inflection was, in 217 00:11:57,600 --> 00:12:00,959 Speaker 1: Layman's terms, intended to be a more emotion aware version 218 00:12:00,960 --> 00:12:04,360 Speaker 1: of CHATGBT just one year after its creation, they'd raised 219 00:12:04,400 --> 00:12:07,360 Speaker 1: one point three billion dollars at evaluation of four billion. 220 00:12:07,920 --> 00:12:11,400 Speaker 1: As I noted previously, I believe that Microsoft's acquisition of 221 00:12:11,400 --> 00:12:14,800 Speaker 1: Inflection was an attempt to avoid regulatory scrutiny by structuring 222 00:12:14,800 --> 00:12:17,600 Speaker 1: an acquisition as a transfer of tech and talent, rather 223 00:12:17,640 --> 00:12:21,480 Speaker 1: than buying the company where the tech and the talent live. 224 00:12:22,040 --> 00:12:24,599 Speaker 1: And yet I'm also convinced that Inflection would struggle to 225 00:12:24,679 --> 00:12:27,440 Speaker 1: last out on its own, even with its smaller user 226 00:12:27,480 --> 00:12:29,960 Speaker 1: base and with a one point three billion dollar war chest, 227 00:12:29,960 --> 00:12:32,640 Speaker 1: and eventually it would have had to get acquired or 228 00:12:32,760 --> 00:12:37,679 Speaker 1: just die. Regardless, I'm beginning to see it patten AI's 229 00:12:37,720 --> 00:12:42,280 Speaker 1: point of failure is shifting, or more accurately centralizing. At 230 00:12:42,280 --> 00:12:46,120 Speaker 1: one point, the burden of this entire generative AI fast 231 00:12:46,240 --> 00:12:48,840 Speaker 1: was shouldered by a large and disparate group of startups 232 00:12:48,840 --> 00:12:51,840 Speaker 1: and investors, and it's now moving to a few shoulders, 233 00:12:51,880 --> 00:12:55,000 Speaker 1: those of giants like Microsoft, Google, Amazon, Meta and to 234 00:12:55,120 --> 00:12:59,080 Speaker 1: a much lesser extent, Apple. The two most prevalent large 235 00:12:59,160 --> 00:13:02,840 Speaker 1: language models outside of course of Metaslama open AI's GPT 236 00:13:02,960 --> 00:13:06,880 Speaker 1: and Anthropics claud are effectively big tech welfare recipients receiving 237 00:13:06,920 --> 00:13:09,400 Speaker 1: billions of dollars in cloud credits to run their extremely 238 00:13:09,440 --> 00:13:12,839 Speaker 1: expensive models without having to build their own infrastructure. It's 239 00:13:12,840 --> 00:13:15,960 Speaker 1: also kind of a con because that's not real money, 240 00:13:16,240 --> 00:13:19,679 Speaker 1: that's chuckie geese tokens, and how are those valued. Ten 241 00:13:19,720 --> 00:13:22,559 Speaker 1: billion dollars goes to open Ai, mostly in cloud credits, 242 00:13:22,640 --> 00:13:25,480 Speaker 1: and they get equity in return. That doesn't make it. 243 00:13:25,520 --> 00:13:27,800 Speaker 1: I can't invest in companies and air miles. Why can 244 00:13:27,840 --> 00:13:32,280 Speaker 1: Microsoft do it? Anyway? Microsoft invested billions of dollars in 245 00:13:32,320 --> 00:13:35,360 Speaker 1: open Ai, and Google and Amazon are propping Upanthropic too, 246 00:13:35,400 --> 00:13:37,800 Speaker 1: with the latter completing its four billion dollar investment in 247 00:13:37,800 --> 00:13:40,880 Speaker 1: the ai company earlier this year. I need to be 248 00:13:41,000 --> 00:13:43,760 Speaker 1: clear that open Ai and Anthropic are wholly reliant on 249 00:13:43,840 --> 00:13:46,480 Speaker 1: the mercies of a handful of trillion dollar companies who 250 00:13:46,600 --> 00:13:49,280 Speaker 1: are themselves at the mercy of the rot economy in 251 00:13:49,320 --> 00:13:52,440 Speaker 1: the public markets, and especially in the case of open 252 00:13:52,480 --> 00:13:55,520 Speaker 1: AI's relationship with Microsoft, they're not just dependent on big 253 00:13:55,559 --> 00:13:58,240 Speaker 1: tech firms financial support, but also on their ability to 254 00:13:58,280 --> 00:14:01,679 Speaker 1: procure and build the actual infratre structure of these products. 255 00:14:02,320 --> 00:14:05,640 Speaker 1: Open Ai does not have their own service, neither does Anthropic. 256 00:14:06,000 --> 00:14:09,160 Speaker 1: Perhaps they have some a few, but the majority of 257 00:14:09,240 --> 00:14:12,760 Speaker 1: their actual processing is done by big tech firms that 258 00:14:13,600 --> 00:14:16,840 Speaker 1: paid them in Chuck e GE's tokens in cloud credits 259 00:14:16,840 --> 00:14:19,920 Speaker 1: to run their services. Who knows how you even amortize that. 260 00:14:20,120 --> 00:14:23,600 Speaker 1: It's completely insane. It drives me insane that they're able 261 00:14:23,680 --> 00:14:27,080 Speaker 1: to do this every time they find a horrible little 262 00:14:27,080 --> 00:14:30,800 Speaker 1: thing to do. But me, I'm at the mercy of 263 00:14:30,840 --> 00:14:32,840 Speaker 1: one of the fine products that you'll hear after this 264 00:14:32,920 --> 00:14:36,280 Speaker 1: ad break, one that will no doubt echo my exact thoughts, feelings, 265 00:14:36,280 --> 00:14:39,560 Speaker 1: and desires. Buy their things or don't. That's up to you, 266 00:14:49,760 --> 00:14:54,720 Speaker 1: and we're back. As I was previously saying, Friday, August second, 267 00:14:54,840 --> 00:14:56,960 Speaker 1: twenty twenty four is a rough day for AI, and 268 00:14:57,000 --> 00:15:01,720 Speaker 1: it somehow got worse. That's SAE. The Information reported that 269 00:15:01,840 --> 00:15:05,840 Speaker 1: Nvidia has reportedly told Microsoft and another unnamed cloud customer 270 00:15:06,040 --> 00:15:09,560 Speaker 1: that its next generation Blackwell chips, which are designed primarily 271 00:15:09,560 --> 00:15:12,760 Speaker 1: for accelerating AI compute tasks, will be delayed by three 272 00:15:12,800 --> 00:15:16,400 Speaker 1: months due to an unspecified design floor identified by its 273 00:15:16,400 --> 00:15:21,800 Speaker 1: contract foundry TSMC. It's entirely possible that this floor takes 274 00:15:21,840 --> 00:15:25,640 Speaker 1: a little more than three months to resolve. Semiconductor manufacturing 275 00:15:25,720 --> 00:15:28,880 Speaker 1: is really, really, really really hard and issues that initially 276 00:15:28,880 --> 00:15:32,120 Speaker 1: see minor can rapidly spiral out of control. On top 277 00:15:32,160 --> 00:15:34,840 Speaker 1: of that, you can't rush them in the same way 278 00:15:34,880 --> 00:15:37,600 Speaker 1: you can with software. You can't just throw a bunch 279 00:15:37,680 --> 00:15:41,320 Speaker 1: more money. TSMC has plenty. But on top of that, 280 00:15:42,240 --> 00:15:45,360 Speaker 1: this isn't a money problem. This is one of those 281 00:15:45,440 --> 00:15:48,960 Speaker 1: scientific ones that you can't just skip over. And with 282 00:15:49,000 --> 00:15:52,000 Speaker 1: black Whill representing a major leap forward in capabilities both 283 00:15:52,000 --> 00:15:54,720 Speaker 1: in the terms of power efficiency and sheer compute power, 284 00:15:55,280 --> 00:15:57,440 Speaker 1: this is a huge blow to open ai and it's 285 00:15:57,640 --> 00:16:01,360 Speaker 1: many many competitors, by which I mean it's other competitor, 286 00:16:01,400 --> 00:16:05,280 Speaker 1: which is also bankrolled by a big tech sugar daddy. 287 00:16:05,320 --> 00:16:07,840 Speaker 1: It's also likely to delay some of open AI's grander 288 00:16:07,840 --> 00:16:10,800 Speaker 1: ambitions for an indeterminate amount of time, which in turn 289 00:16:11,280 --> 00:16:14,200 Speaker 1: kind of dampens its appeal to potential investors. And to 290 00:16:14,240 --> 00:16:16,760 Speaker 1: be clear, open ai basically has to raise in the 291 00:16:16,800 --> 00:16:19,920 Speaker 1: next six months, as the Information reported that open ai 292 00:16:20,000 --> 00:16:23,440 Speaker 1: could lose as much as five billion dollars this year alone. 293 00:16:23,640 --> 00:16:26,320 Speaker 1: I will get to that later, but just to be clear, 294 00:16:27,000 --> 00:16:30,200 Speaker 1: that's not five billion in losses plus revenue. That is, 295 00:16:31,560 --> 00:16:34,960 Speaker 1: after you take the money out that they made. It's 296 00:16:35,040 --> 00:16:38,200 Speaker 1: really bad it's really really bad. Though. It's all so 297 00:16:38,480 --> 00:16:41,400 Speaker 1: very bad. I'm not sure why more people aren't freaking out. 298 00:16:42,240 --> 00:16:45,240 Speaker 1: I'm mostly chilling just because they don't really care if 299 00:16:45,240 --> 00:16:47,920 Speaker 1: they live or die. I feel bad for the people 300 00:16:47,960 --> 00:16:50,680 Speaker 1: with the jobs, but I'm rambling. We'll get back to it, 301 00:16:51,440 --> 00:16:53,520 Speaker 1: and this is a problem that expands to Google and 302 00:16:53,560 --> 00:16:56,760 Speaker 1: Microsoft too. In Nvidious Blackwell chips was supposed to be 303 00:16:56,760 --> 00:17:00,320 Speaker 1: a major technological leap for generative AI, designed with both 304 00:17:00,320 --> 00:17:03,160 Speaker 1: influence the way that large language models generate answers in 305 00:17:03,200 --> 00:17:06,240 Speaker 1: training in mind, and thus providing vastly more compute parerent 306 00:17:06,400 --> 00:17:10,480 Speaker 1: energy efficiency necessary to make this shit move. But as 307 00:17:10,520 --> 00:17:13,280 Speaker 1: of now, these chips have been delayed, with their availability 308 00:17:13,359 --> 00:17:15,800 Speaker 1: not expected into the first quarter of twenty twenty five 309 00:17:16,040 --> 00:17:19,200 Speaker 1: at the earliest, and the Information reports that open Ai 310 00:17:19,280 --> 00:17:21,920 Speaker 1: probably won't get access to them until March twenty twenty four, 311 00:17:22,760 --> 00:17:25,320 Speaker 1: and Nvidia might be forced to do new test runs 312 00:17:25,359 --> 00:17:27,879 Speaker 1: with its foundry partner before scaling up to mass production. 313 00:17:29,000 --> 00:17:32,040 Speaker 1: One particularly worrying quote, and I kind of just hinted 314 00:17:32,040 --> 00:17:35,000 Speaker 1: in it there, was that Microsoft managers had planned to 315 00:17:35,000 --> 00:17:38,040 Speaker 1: make Blackwell powered servers available to open Ai by January, 316 00:17:38,080 --> 00:17:40,320 Speaker 1: but like I said, may need to plan for March 317 00:17:40,359 --> 00:17:42,520 Speaker 1: or early spring, and that was said by a person 318 00:17:42,560 --> 00:17:47,439 Speaker 1: with knowledge of the situation. Beautiful name the reason that's bad, Well, 319 00:17:47,920 --> 00:17:52,359 Speaker 1: it's not really just one. First, open ai desperately needs 320 00:17:52,400 --> 00:17:55,320 Speaker 1: something new. They need something that will show both investors 321 00:17:55,320 --> 00:17:58,359 Speaker 1: and the media that open ai is building something meaningful. 322 00:17:58,560 --> 00:18:00,720 Speaker 1: And while it's possible it will be able to deliver 323 00:18:00,800 --> 00:18:03,240 Speaker 1: GPT five, which is their next model, in the near 324 00:18:03,320 --> 00:18:06,640 Speaker 1: term future, even potential customers don't believe the jump will 325 00:18:06,680 --> 00:18:10,800 Speaker 1: be significant enough from GPT four. Second, for reasons I've 326 00:18:10,840 --> 00:18:13,600 Speaker 1: alluded to at the start of this episode, investors are 327 00:18:13,640 --> 00:18:16,439 Speaker 1: getting a touch itchy with the AI hype boom, and 328 00:18:16,560 --> 00:18:19,639 Speaker 1: more specifically with the giant tech companies that are bankrolling it. 329 00:18:20,200 --> 00:18:22,840 Speaker 1: The street needs to see results, or at the very 330 00:18:22,920 --> 00:18:26,280 Speaker 1: least ap plausible case for the future profitability and marketability 331 00:18:26,440 --> 00:18:29,520 Speaker 1: of generative AI, a thing that it's genuinely bad at. 332 00:18:30,320 --> 00:18:33,040 Speaker 1: In vidious Blackwell chips could help open ai models run 333 00:18:33,119 --> 00:18:36,880 Speaker 1: faster and ingest training data faster, as well as, even 334 00:18:36,960 --> 00:18:40,320 Speaker 1: if it's not possible, fooling investors into believing that open 335 00:18:40,359 --> 00:18:42,520 Speaker 1: ai had the latest tech and thus could build the 336 00:18:42,960 --> 00:18:46,119 Speaker 1: average general intelligence that sam Ortmand's been lying about and 337 00:18:46,320 --> 00:18:51,600 Speaker 1: allow him to continue his con regardless, the article from 338 00:18:51,600 --> 00:18:54,720 Speaker 1: the Information put it very plainly open AI only gets 339 00:18:54,760 --> 00:18:57,960 Speaker 1: access to the latest technology as fast as Microsoft permits, 340 00:18:58,080 --> 00:19:00,199 Speaker 1: or as fast as Microsoft is able to if I 341 00:19:00,240 --> 00:19:03,480 Speaker 1: did on their own, even one of these events would 342 00:19:03,480 --> 00:19:06,040 Speaker 1: be a deeply worrying sign of the bubble popping, and 343 00:19:06,080 --> 00:19:08,080 Speaker 1: together they threatened to begin a collapse that I've been 344 00:19:08,080 --> 00:19:11,000 Speaker 1: predicting since March, where I said that AI companies had 345 00:19:11,000 --> 00:19:14,320 Speaker 1: about three quarters to prove themselves and I quote savaging 346 00:19:14,320 --> 00:19:17,000 Speaker 1: the revenues of the biggest companies in tech when things 347 00:19:17,040 --> 00:19:20,960 Speaker 1: don't work out. But I've realized now that it isn't 348 00:19:20,960 --> 00:19:24,000 Speaker 1: really super useful to attach things to time. Though I 349 00:19:24,160 --> 00:19:26,439 Speaker 1: stand by bud prediction, and thus I think it's more 350 00:19:26,480 --> 00:19:29,159 Speaker 1: illustrative to suggest what the terms of the bubble popping 351 00:19:29,160 --> 00:19:33,880 Speaker 1: actually are, So let's define them. What do I mean 352 00:19:33,880 --> 00:19:37,120 Speaker 1: when I talk about the bubble popping? For the sake 353 00:19:37,160 --> 00:19:40,080 Speaker 1: of clarity, I'm defining it as the major cloud companies 354 00:19:40,119 --> 00:19:43,159 Speaker 1: reducing capital expenditures related to generative AI in a public 355 00:19:43,200 --> 00:19:46,520 Speaker 1: and significant manner, as in an actual statement from satching 356 00:19:46,560 --> 00:19:49,760 Speaker 1: Adelaer at Microsoft or Sun Duppishi at Google or one 357 00:19:49,760 --> 00:19:52,640 Speaker 1: of the major large language model companies Anthropic or open 358 00:19:52,680 --> 00:19:56,880 Speaker 1: AI collapsing in some way. No cheap thrills, no half measures. 359 00:19:57,280 --> 00:20:00,760 Speaker 1: We're not calling this fucking thing until it actually pops, 360 00:20:01,000 --> 00:20:03,080 Speaker 1: and it will not pop into one or both of 361 00:20:03,119 --> 00:20:06,119 Speaker 1: these companies go tits up. To be clear, when I 362 00:20:06,160 --> 00:20:10,040 Speaker 1: talk about Anthropic and Open Air collapsing, I don't actually 363 00:20:10,119 --> 00:20:13,760 Speaker 1: mean in the Enron style way. These companies are probably 364 00:20:13,800 --> 00:20:18,240 Speaker 1: not going to fall apart, shut down, close their doors suddenly. 365 00:20:19,320 --> 00:20:22,439 Speaker 1: It could be that they're absorbed into another company on 366 00:20:22,520 --> 00:20:26,800 Speaker 1: unfavorable terms, you know, like character AI and inflection, like 367 00:20:26,880 --> 00:20:29,800 Speaker 1: the things that just happened. And in that case, I 368 00:20:29,800 --> 00:20:32,840 Speaker 1: think investors are going to take a massive loss. Open 369 00:20:32,880 --> 00:20:36,040 Speaker 1: AI or Anthropic might be forced to radically limit their operations, 370 00:20:36,080 --> 00:20:39,359 Speaker 1: either by shutting down their free chatbots or limiting access 371 00:20:39,359 --> 00:20:42,520 Speaker 1: to enterprise customers, or by slowing down development massively, which 372 00:20:42,640 --> 00:20:46,200 Speaker 1: by the way, will be death's embrace. It will be over. 373 00:20:47,000 --> 00:20:49,000 Speaker 1: We could see massive layoffs, or we could see these 374 00:20:49,000 --> 00:20:51,800 Speaker 1: companies pivot to a less capital intensive business model, such 375 00:20:51,800 --> 00:20:55,800 Speaker 1: as licensing their patterns and IP to other technologies or companies, 376 00:20:56,200 --> 00:20:58,960 Speaker 1: but not really offering an actual tangible product to consumers 377 00:20:59,040 --> 00:21:01,400 Speaker 1: or business customers, as opposed to what they're doing today, 378 00:21:01,400 --> 00:21:04,800 Speaker 1: which is offering a tangible thing that doesn't really doom 379 00:21:04,840 --> 00:21:08,640 Speaker 1: that much. I'm rambling again. Yeah, I can imagine all 380 00:21:08,680 --> 00:21:11,280 Speaker 1: sorts of terminal endings for Generative AI is big two, 381 00:21:11,280 --> 00:21:13,840 Speaker 1: and some of them aren't even legally actionable. But I 382 00:21:13,880 --> 00:21:16,840 Speaker 1: believe that the collapse or absorption of these companies is 383 00:21:16,880 --> 00:21:20,160 Speaker 1: the one critical sign to look for. Open ai began 384 00:21:20,240 --> 00:21:23,320 Speaker 1: this hype cycle and Sam Altman is definitely the pt 385 00:21:23,480 --> 00:21:26,440 Speaker 1: barnum of the large language model circus, and open ai 386 00:21:26,520 --> 00:21:29,000 Speaker 1: has absorbed more money and attention than any other startup 387 00:21:29,040 --> 00:21:30,879 Speaker 1: of the last few years, and outside of Uber, I 388 00:21:30,920 --> 00:21:34,520 Speaker 1: think they might be getting the most all time. God, 389 00:21:34,560 --> 00:21:37,000 Speaker 1: I hate it. It's symbolic of the excess in the 390 00:21:37,000 --> 00:21:39,199 Speaker 1: waste of the generative AI boom and it's death or 391 00:21:39,320 --> 00:21:42,960 Speaker 1: as I mentioned, some alternative collapse is the sign that 392 00:21:42,960 --> 00:21:45,240 Speaker 1: we're done here in the same way that FTX signaled 393 00:21:45,240 --> 00:21:48,600 Speaker 1: the end of the cryptocurrency boom. And if open AI's 394 00:21:48,680 --> 00:21:52,000 Speaker 1: collapse marks the apocalypse for generative AI, I believe it 395 00:21:52,040 --> 00:21:54,800 Speaker 1: will be accompanied by one or several of the following 396 00:21:54,840 --> 00:21:58,560 Speaker 1: pale horses, real tangible signs that things are falling apart, 397 00:21:58,760 --> 00:22:01,160 Speaker 1: things you can send to your friends, family and laugh at. 398 00:22:02,040 --> 00:22:06,399 Speaker 1: Let's take a look, shall we. Number one, Any significant 399 00:22:06,400 --> 00:22:09,000 Speaker 1: price changes by any large language model company are a 400 00:22:09,000 --> 00:22:12,920 Speaker 1: bad sign. Desperate times require desperate measures, and any oscillation 401 00:22:13,040 --> 00:22:15,119 Speaker 1: in pricing is a bad sign. We could see a 402 00:22:15,160 --> 00:22:17,080 Speaker 1: race to the bottom, which is kind of already happening 403 00:22:17,080 --> 00:22:20,960 Speaker 1: with model developers releasing increasingly cheaper options to grow revenue 404 00:22:20,960 --> 00:22:23,960 Speaker 1: and build their customer bases, which is an idea that 405 00:22:24,000 --> 00:22:26,680 Speaker 1: only makes sense if the underpinning technology is profitable, which 406 00:22:26,760 --> 00:22:29,960 Speaker 1: Generative Ai is not, and it's already kind of begun 407 00:22:30,040 --> 00:22:33,119 Speaker 1: with open AI's reduced price GPT four oh mini in 408 00:22:33,240 --> 00:22:37,840 Speaker 1: Google Gemini's Google Flash one point five I believe it's called, 409 00:22:37,880 --> 00:22:40,360 Speaker 1: and they just reduce that pricing to compete with GPT 410 00:22:40,480 --> 00:22:44,560 Speaker 1: four oh mini's cheapest option. This is totally unsustainable, which 411 00:22:44,600 --> 00:22:46,280 Speaker 1: means it fits right in with the rest of the 412 00:22:46,359 --> 00:22:50,879 Speaker 1: hype cycle. Conversely, if prices start increasing, this is a 413 00:22:50,920 --> 00:22:53,399 Speaker 1: sign that the company is getting really, really desperate and 414 00:22:53,440 --> 00:22:55,640 Speaker 1: have to find a way to start recouping the massive 415 00:22:55,680 --> 00:23:00,240 Speaker 1: costs of the unsustainable shit show they've been participating in well. 416 00:23:00,280 --> 00:23:02,520 Speaker 1: This I think is probably one of the later ones. 417 00:23:02,600 --> 00:23:05,120 Speaker 1: If they do this, if they even get to it, 418 00:23:05,119 --> 00:23:07,879 Speaker 1: it's just the sign that things are falling apart. I 419 00:23:07,920 --> 00:23:10,520 Speaker 1: also think we're going to see stories about general discord 420 00:23:10,560 --> 00:23:13,840 Speaker 1: in AI investment. Any articles you see about investors fleeing 421 00:23:13,960 --> 00:23:16,760 Speaker 1: private aideals as they did with the metaverse, or a 422 00:23:16,800 --> 00:23:19,400 Speaker 1: sign that things are falling apart. And the things I'm hearing, 423 00:23:19,440 --> 00:23:21,399 Speaker 1: by the way, is that this has already begun, but 424 00:23:21,480 --> 00:23:24,640 Speaker 1: that's anecdotal. Don't take my word for it. Just watch 425 00:23:24,680 --> 00:23:27,840 Speaker 1: the media. We might also see stories about open ai 426 00:23:28,000 --> 00:23:31,119 Speaker 1: having trouble raising money. Up until now, there haven't been 427 00:23:31,160 --> 00:23:33,720 Speaker 1: any rumors about them having trouble, despite the fact that 428 00:23:33,920 --> 00:23:36,440 Speaker 1: they're expected to lose like I mentioned, five billion dollars 429 00:23:36,520 --> 00:23:39,479 Speaker 1: in twenty twenty four. They need to raise billions of 430 00:23:39,520 --> 00:23:41,560 Speaker 1: dollars and they have to do it soon. And if 431 00:23:41,600 --> 00:23:45,920 Speaker 1: they can't raise, nobody can. But moving on, there's another 432 00:23:45,960 --> 00:23:48,320 Speaker 1: harbinger to look out for, and it would be evidence 433 00:23:48,359 --> 00:23:51,520 Speaker 1: that Google or Microsoft is reducing the capex as discussed, 434 00:23:52,359 --> 00:23:55,600 Speaker 1: because it's really important to realize that venture capital and 435 00:23:55,640 --> 00:23:58,600 Speaker 1: Silicon Valley really are not They're not the ones propping 436 00:23:58,640 --> 00:24:00,840 Speaker 1: this up. They might be the year leaders, they might 437 00:24:00,840 --> 00:24:03,720 Speaker 1: be the ones talking about how important is But Google 438 00:24:03,760 --> 00:24:06,440 Speaker 1: and Microsoft are the ones holding this up, and Amazon 439 00:24:06,480 --> 00:24:09,240 Speaker 1: as well. They're the ones who are bankrolling Anthropic and 440 00:24:09,280 --> 00:24:12,560 Speaker 1: open Ai. Microsoft as I will get to basically owns 441 00:24:12,560 --> 00:24:15,560 Speaker 1: open ai. If those companies decide they don't want this 442 00:24:15,640 --> 00:24:20,879 Speaker 1: to continue, it will stop. Venture capital cannot afford to 443 00:24:20,960 --> 00:24:24,440 Speaker 1: keep open ai alive. But moving on, you should keep 444 00:24:24,480 --> 00:24:28,119 Speaker 1: out and beady little I for any discord within any 445 00:24:28,160 --> 00:24:30,679 Speaker 1: of the major players in generative AI. But I'm not 446 00:24:30,720 --> 00:24:33,639 Speaker 1: talking about open ai and Anthropic, talking about companies like 447 00:24:33,680 --> 00:24:36,560 Speaker 1: Scale Ai, who are training data company that's raised over 448 00:24:36,760 --> 00:24:39,240 Speaker 1: one on a half billion dollars and cohere, another large 449 00:24:39,320 --> 00:24:41,800 Speaker 1: language model company that recently raised another four hundred and 450 00:24:41,840 --> 00:24:45,639 Speaker 1: fifty million dollars from Nvidia and Salesforce and Salesforce up. 451 00:24:46,320 --> 00:24:48,359 Speaker 1: They've been talking about doing AI for like a decade. 452 00:24:48,600 --> 00:24:52,080 Speaker 1: I don't know what product they sell. I'd watch those 453 00:24:52,119 --> 00:24:56,000 Speaker 1: two very carefully. They're seen as stable, revenue generating yet 454 00:24:56,119 --> 00:24:59,640 Speaker 1: as discussed, unprofitable and sillery firms that will likely need 455 00:24:59,680 --> 00:25:02,280 Speaker 1: to raise again in the next twelve months. If we 456 00:25:02,359 --> 00:25:05,560 Speaker 1: hear about problems such as employees being unhappy, or trouble's 457 00:25:05,560 --> 00:25:08,520 Speaker 1: fundraising or making money. This means that people on the 458 00:25:08,520 --> 00:25:11,280 Speaker 1: inside are making nasty little noises because they know things 459 00:25:11,320 --> 00:25:14,280 Speaker 1: are falling apart. And the same, by the way, goes 460 00:25:14,280 --> 00:25:17,840 Speaker 1: for any discord in open Ai or Anthropic. This one's obvious. 461 00:25:18,160 --> 00:25:20,520 Speaker 1: If we hear there are problems they're worried, or if 462 00:25:20,560 --> 00:25:24,160 Speaker 1: they're getting worried, if they can't raise money, if they're 463 00:25:24,200 --> 00:25:27,960 Speaker 1: losing people, which is kind of already happening, if they 464 00:25:28,000 --> 00:25:29,920 Speaker 1: can't seem to make a deal work, if they're having 465 00:25:30,000 --> 00:25:35,080 Speaker 1: trouble progressing their software. Anything about these companies is probably 466 00:25:35,080 --> 00:25:37,520 Speaker 1: a bad sign. If I were you, I'd keep an 467 00:25:37,560 --> 00:25:40,840 Speaker 1: eye on the Washington Post. Natasha Tiku's done great work there, 468 00:25:41,000 --> 00:25:43,639 Speaker 1: as the people at Shera A. V. Day and Jeffrey Fowler. 469 00:25:43,680 --> 00:25:47,240 Speaker 1: But Natasha's done some great work. Or mister Altman, the 470 00:25:47,280 --> 00:25:49,560 Speaker 1: Information has got some great people doing work on this 471 00:25:49,760 --> 00:25:52,400 Speaker 1: Reuter's Tech Crunch for or film media. They're the ones 472 00:25:52,400 --> 00:25:56,280 Speaker 1: that keep an eye out for. We've already seen a 473 00:25:56,280 --> 00:25:59,880 Speaker 1: little bit of this happening already. A few days ago, 474 00:26:00,080 --> 00:26:03,000 Speaker 1: open Ai lost two of its co founders, John Shulman, 475 00:26:03,080 --> 00:26:06,280 Speaker 1: left for competitor anthropic and chief operating officer Greg Brockman 476 00:26:06,560 --> 00:26:10,480 Speaker 1: has gone on something he's calling extended leave. As on aside, 477 00:26:10,520 --> 00:26:13,119 Speaker 1: this is objectively one of the funniest things it's anyone 478 00:26:13,160 --> 00:26:15,920 Speaker 1: has ever done in tech. Yeah, man, nothing big going on. 479 00:26:16,640 --> 00:26:19,439 Speaker 1: Enjoy a vacation, mate, Not like the entire industry is 480 00:26:19,480 --> 00:26:22,439 Speaker 1: falling apart, not like the bubble might be popping. Enjoy 481 00:26:22,480 --> 00:26:26,520 Speaker 1: carbo Anyway. As things get especially bad, I think we're 482 00:26:26,520 --> 00:26:29,640 Speaker 1: going to see discussions of layoffs based on people I've 483 00:26:29,680 --> 00:26:33,400 Speaker 1: talked to. Generative AI companies are paying these massive salaries 484 00:26:33,440 --> 00:26:37,080 Speaker 1: in stock and cash, and they're not run particularly efficiently 485 00:26:37,160 --> 00:26:40,320 Speaker 1: because they are really driven by dreamers, by which I 486 00:26:40,359 --> 00:26:43,359 Speaker 1: mean conn artists. If AI jobs are no longer these 487 00:26:43,400 --> 00:26:46,000 Speaker 1: cushy little Heidi holes for the most expensive engineers in 488 00:26:46,040 --> 00:26:48,760 Speaker 1: the valley, things are going to sour fast because they 489 00:26:48,800 --> 00:26:51,240 Speaker 1: need the best to do this mediocre crab that you 490 00:26:51,320 --> 00:26:54,359 Speaker 1: see every day. But I'd argue the biggest and most 491 00:26:54,400 --> 00:26:57,159 Speaker 1: desperate of these pale horses would be some sort of big, 492 00:26:57,480 --> 00:27:03,200 Speaker 1: stupid magic trick. As these companies get desperate, expect someone, 493 00:27:03,400 --> 00:27:06,000 Speaker 1: especially Open AI, to try and show something new and 494 00:27:06,119 --> 00:27:09,320 Speaker 1: crazy as a means of trying to turn this narrative around. 495 00:27:09,560 --> 00:27:12,840 Speaker 1: When or if this happens, look very carefully at what 496 00:27:12,880 --> 00:27:15,919 Speaker 1: they say about the products availability, what it can actually do, 497 00:27:16,240 --> 00:27:18,840 Speaker 1: or who they show it to. Is it available publicly? 498 00:27:19,080 --> 00:27:22,280 Speaker 1: Are there limitations on it? And are there any weird 499 00:27:22,359 --> 00:27:25,560 Speaker 1: terms and conditions to look out for. If open ai 500 00:27:25,720 --> 00:27:27,960 Speaker 1: gets desperate, it may move up the public launch of 501 00:27:28,000 --> 00:27:31,399 Speaker 1: soro It's generative video product. Doing those really only going 502 00:27:31,480 --> 00:27:33,880 Speaker 1: to cause more problems. There isn't a chance in hell 503 00:27:34,240 --> 00:27:37,159 Speaker 1: that this thing is profitable, and I'm one hundred percent 504 00:27:37,200 --> 00:27:40,119 Speaker 1: confident that it's much much, much, much more expensive than 505 00:27:40,160 --> 00:27:43,960 Speaker 1: the video and the text based generation tools they already have, 506 00:27:44,400 --> 00:27:47,760 Speaker 1: and I imagine its visual ginconsistencies and hallucinations would make 507 00:27:47,800 --> 00:27:51,480 Speaker 1: for some entertaining content for YouTubers and tech reporters, which 508 00:27:51,520 --> 00:27:55,560 Speaker 1: will mean that they just have hundreds thousands, hundreds of thousands, 509 00:27:55,560 --> 00:27:58,560 Speaker 1: perhaps the people requesting shitty videos off of this thing, 510 00:27:58,720 --> 00:28:02,800 Speaker 1: burning servers quite literally in this case, losing the money. 511 00:28:02,880 --> 00:28:05,200 Speaker 1: And for what how much can they charge for this 512 00:28:05,240 --> 00:28:09,240 Speaker 1: thing because it's not going to be enough. The common 513 00:28:09,280 --> 00:28:11,320 Speaker 1: thread between all of these events is that they're all 514 00:28:11,320 --> 00:28:15,359 Speaker 1: expressions of desperation fear and a total lack of confidence 515 00:28:15,400 --> 00:28:19,200 Speaker 1: in the underlying company. So far, and this is Napkin maths, 516 00:28:19,240 --> 00:28:21,480 Speaker 1: I'd estimate that a total of two hundred billion dollars 517 00:28:21,560 --> 00:28:23,919 Speaker 1: has been spent to get Generative AI to this point 518 00:28:24,160 --> 00:28:27,199 Speaker 1: in infrastructure and funding and energy in so many different 519 00:28:27,280 --> 00:28:29,440 Speaker 1: meaningless ways, or to get us to the point that 520 00:28:29,480 --> 00:28:31,240 Speaker 1: we have a tool that's really good at generating things 521 00:28:31,240 --> 00:28:33,040 Speaker 1: that aren't as good as what a humor could make, 522 00:28:33,680 --> 00:28:36,440 Speaker 1: and only sometimes are they good enough to actually use. 523 00:28:37,320 --> 00:28:40,200 Speaker 1: To be clear, we're not quite at the bubble popping yet. 524 00:28:40,440 --> 00:28:42,680 Speaker 1: The reason I'd chose the collapse of open ai as 525 00:28:42,840 --> 00:28:46,040 Speaker 1: the event is because it will mean that Microsoft decided 526 00:28:46,040 --> 00:28:48,600 Speaker 1: to cut them off and that there wasn't enough banker 527 00:28:48,680 --> 00:28:51,920 Speaker 1: or venture capital interest to prop them up further. Open 528 00:28:51,960 --> 00:28:55,040 Speaker 1: AI's collapse would be both financial and symbolic, the sign 529 00:28:55,080 --> 00:28:57,440 Speaker 1: that the valley would let a company die and that 530 00:28:57,520 --> 00:28:59,680 Speaker 1: this idea was not good enough for everybody to stake 531 00:28:59,680 --> 00:29:03,440 Speaker 1: their fetes on it. In any case, things could change, 532 00:29:03,480 --> 00:29:07,400 Speaker 1: the bubble could stay inflated. But while you're thinking about how, 533 00:29:07,560 --> 00:29:09,520 Speaker 1: why don't you take a breather to hear from someone 534 00:29:10,120 --> 00:29:14,640 Speaker 1: one of our wonderful beautiful, precisely chosen advertisers that you 535 00:29:14,720 --> 00:29:31,000 Speaker 1: simply must pay money to, and we're back now. I 536 00:29:31,000 --> 00:29:32,600 Speaker 1: don't want to give you any homework. I no, you're 537 00:29:32,600 --> 00:29:35,000 Speaker 1: currently holding up a convenience store or maybe trying to 538 00:29:35,000 --> 00:29:37,520 Speaker 1: get the hub cap off of a car wheel. You 539 00:29:37,600 --> 00:29:39,680 Speaker 1: hear me speak enough. But a few weeks ago I 540 00:29:39,680 --> 00:29:42,240 Speaker 1: wrote a newsletter called how does open ai Survive? And 541 00:29:42,320 --> 00:29:44,960 Speaker 1: eight thousand or so word analysis of how difficult it's 542 00:29:45,000 --> 00:29:47,160 Speaker 1: going to be to keep that company going at its 543 00:29:47,200 --> 00:29:50,400 Speaker 1: current burn rate, to simmer it down as quickly as possible. 544 00:29:50,520 --> 00:29:53,200 Speaker 1: Open ai costs about eight point five billion dollars a 545 00:29:53,240 --> 00:29:55,120 Speaker 1: year to run and only makes about three and a 546 00:29:55,160 --> 00:29:56,960 Speaker 1: half billion to four and a half billion a year 547 00:29:56,960 --> 00:29:59,200 Speaker 1: in revenue, a number, by the way that I'm that 548 00:29:59,280 --> 00:30:02,520 Speaker 1: does not sound right, but it's what's being reported. Open 549 00:30:02,560 --> 00:30:05,760 Speaker 1: AI's costs are increasingly nearly and generative AI, as I've 550 00:30:05,800 --> 00:30:09,840 Speaker 1: said a few times, is deeply unprofitable and unsustainable. Making 551 00:30:09,880 --> 00:30:13,880 Speaker 1: open AI's models better, which doesn't necessarily mean they're more 552 00:30:13,960 --> 00:30:17,400 Speaker 1: valuable and more useful, also costs hundreds of millions, if 553 00:30:17,440 --> 00:30:21,040 Speaker 1: not billions, of dollars. As a result, open ai needs 554 00:30:21,040 --> 00:30:23,520 Speaker 1: to raise more money than anybody has ever raised in 555 00:30:23,800 --> 00:30:25,800 Speaker 1: the history of the valley in the next twelve to 556 00:30:25,800 --> 00:30:28,920 Speaker 1: twenty four months, or launch a products so significant that 557 00:30:28,960 --> 00:30:31,760 Speaker 1: it blows chet GPT out of the water and actually 558 00:30:31,840 --> 00:30:36,440 Speaker 1: is profitable. And these things are unlikely but possible. I 559 00:30:36,440 --> 00:30:41,240 Speaker 1: guess Microsoft, Google, and Amazon could somehow change the narrative, 560 00:30:41,280 --> 00:30:44,040 Speaker 1: but doing so too would require some sort of technological 561 00:30:44,040 --> 00:30:47,280 Speaker 1: breakthrough or a product, a very obvious product that would 562 00:30:47,320 --> 00:30:49,880 Speaker 1: make people money, more money than is being spent on 563 00:30:49,960 --> 00:30:54,040 Speaker 1: the thing. And the markets are also capricious, and thus 564 00:30:54,120 --> 00:30:56,880 Speaker 1: I really wouldn't count on there being a narrative turnaround, 565 00:30:57,080 --> 00:31:00,640 Speaker 1: especially with the shifting sands of a vestor sentiment. As 566 00:31:00,640 --> 00:31:03,920 Speaker 1: I've mentioned previously. On August second, Elliott Management, the hedge 567 00:31:03,960 --> 00:31:06,840 Speaker 1: fund with over seventy billion dollars in assets under management, 568 00:31:07,080 --> 00:31:09,600 Speaker 1: said in a letter to clients that AI is overhyped 569 00:31:09,760 --> 00:31:12,440 Speaker 1: and the Nvidia is a bubble, adding that AI had 570 00:31:12,440 --> 00:31:16,760 Speaker 1: not delivered value commensurate with the hype. While it's possible 571 00:31:16,760 --> 00:31:19,600 Speaker 1: there are others that dissent against the rapidly forming consensus 572 00:31:19,640 --> 00:31:22,320 Speaker 1: against generative AI to prop things up, which I do 573 00:31:22,400 --> 00:31:24,840 Speaker 1: think is possible, as especially if you deepen the whole 574 00:31:24,920 --> 00:31:28,800 Speaker 1: with Microsoft or and Video. It's worth putting Elliott Management's 575 00:31:28,840 --> 00:31:33,160 Speaker 1: note in a broader context. It's only the latest voice 576 00:31:33,520 --> 00:31:36,800 Speaker 1: in a loud or at least increasingly louder, chorus of 577 00:31:36,840 --> 00:31:39,760 Speaker 1: critics whose members include esteamed, analyst houses, and some of 578 00:31:39,800 --> 00:31:43,440 Speaker 1: the largest investment banks in the world. Gold and Sax 579 00:31:43,480 --> 00:31:45,200 Speaker 1: put out a report in July that you've heard in 580 00:31:45,200 --> 00:31:48,000 Speaker 1: the pop culture episode that said generative AI was too 581 00:31:48,000 --> 00:31:50,760 Speaker 1: expensive and didn't solve the complex problems that would need 582 00:31:50,800 --> 00:31:54,440 Speaker 1: to justify their expenses. At the end of July, Gartner 583 00:31:54,480 --> 00:31:57,040 Speaker 1: put out a report predicting that thirty percent of generative 584 00:31:57,080 --> 00:32:00,120 Speaker 1: AI products would be abandoned entirely after a proof of 585 00:32:00,200 --> 00:32:03,440 Speaker 1: concept by end of twenty twenty five, and The Washington 586 00:32:03,440 --> 00:32:06,400 Speaker 1: Post reports that Barklay's bank thinks that and I quote 587 00:32:06,680 --> 00:32:09,160 Speaker 1: Wall Street and Lists are expecting big tech companies to 588 00:32:09,160 --> 00:32:11,600 Speaker 1: spend around sixty billion dollars a year on developing AI 589 00:32:11,680 --> 00:32:14,560 Speaker 1: models by twenty twenty six, but reap only twenty billion 590 00:32:14,600 --> 00:32:17,000 Speaker 1: dollars a year in revenue from AI by that point. 591 00:32:17,400 --> 00:32:21,120 Speaker 1: The lastra is gone there for you. That does say revenue, 592 00:32:21,720 --> 00:32:24,480 Speaker 1: that doesn't say profim and by the way, that's only 593 00:32:24,520 --> 00:32:28,040 Speaker 1: a billion dollars more than Microsoft's entire capital expenditures in 594 00:32:28,040 --> 00:32:32,560 Speaker 1: the last quarter alone. Microsoft also has a storied history 595 00:32:32,600 --> 00:32:36,920 Speaker 1: of pumping and dumping ideas for Microsoft. Augmented reality was 596 00:32:36,960 --> 00:32:40,400 Speaker 1: and I quote, an absolute breakthrough in twenty nineteen before 597 00:32:40,480 --> 00:32:42,800 Speaker 1: it was quietly shoved in a corner with mass layoffs 598 00:32:42,800 --> 00:32:45,520 Speaker 1: and a business line enclosures a few years later in 599 00:32:45,560 --> 00:32:49,200 Speaker 1: twenty twenty one, satchur Innidella, their CEO, couldn't overstate how 600 00:32:49,280 --> 00:32:52,360 Speaker 1: much of a breakthrough the metaverse was yet two years later, 601 00:32:52,360 --> 00:32:54,800 Speaker 1: and big props the Preston Growler of Computer World for 602 00:32:54,840 --> 00:32:58,120 Speaker 1: saying this in February twenty twenty three, he laid off 603 00:32:58,120 --> 00:33:02,120 Speaker 1: most of the people involved in everything on artificial intelligence 604 00:33:02,200 --> 00:33:05,920 Speaker 1: and that everything thing that's a headline from Wired magazine. 605 00:33:06,080 --> 00:33:09,960 Speaker 1: Very good. It's also important to remember that Microsoft effectively 606 00:33:10,000 --> 00:33:13,920 Speaker 1: owns open ai as part of their twenty nineteen funding round, 607 00:33:13,920 --> 00:33:16,360 Speaker 1: which I believe was a billion dollars. Microsoft is full 608 00:33:16,440 --> 00:33:19,200 Speaker 1: access to all of open AI's intellectual property and research, 609 00:33:20,000 --> 00:33:24,040 Speaker 1: which makes thorny little question why at this point would 610 00:33:24,040 --> 00:33:26,960 Speaker 1: Microsoft give open ai any more money other than to 611 00:33:27,000 --> 00:33:30,480 Speaker 1: save face. That's a very real possibility that Satch in 612 00:33:30,480 --> 00:33:32,240 Speaker 1: the Dela is surrounded by yes men who don't know 613 00:33:32,280 --> 00:33:35,080 Speaker 1: what they're talking about. Microsoft also has access to sell 614 00:33:35,120 --> 00:33:37,960 Speaker 1: open ais and I quote pre agi products, which is 615 00:33:38,240 --> 00:33:40,800 Speaker 1: all of their products right now, and full access to 616 00:33:40,840 --> 00:33:43,040 Speaker 1: all of their research, as well as and I quote 617 00:33:43,200 --> 00:33:47,600 Speaker 1: certain rights to open AI's intellectual property. Unless open ai 618 00:33:47,720 --> 00:33:50,720 Speaker 1: is capable of delivering something meaningfully different, something that Dela 619 00:33:50,760 --> 00:33:52,920 Speaker 1: can wave in front of analysts faces and show them 620 00:33:53,120 --> 00:33:56,080 Speaker 1: that they can automate all of these jobs away, its 621 00:33:56,080 --> 00:33:58,600 Speaker 1: existence is just another cost center and one that could 622 00:33:58,640 --> 00:34:02,320 Speaker 1: get eliminated in a tough quarter. And to be clear, 623 00:34:02,440 --> 00:34:05,280 Speaker 1: saturn Adella has been extremely frank about his leverage over 624 00:34:05,280 --> 00:34:08,160 Speaker 1: open ai and it's effective ownership of their tech or 625 00:34:08,200 --> 00:34:11,000 Speaker 1: be it buffeted with the usual bullshit platitudes and niceties 626 00:34:11,000 --> 00:34:13,760 Speaker 1: that you'd expect from a shape shifting tech CEO. Monster. 627 00:34:14,480 --> 00:34:17,920 Speaker 1: In a late twenty twenty interview with PR person Kara Swisher, 628 00:34:18,160 --> 00:34:21,239 Speaker 1: sach an Adella said, if open ai disappeared tomorrow, I 629 00:34:21,280 --> 00:34:23,600 Speaker 1: don't want any customer of ours to be worried about it, 630 00:34:23,680 --> 00:34:26,240 Speaker 1: quite honestly, because we have all the rights to continue 631 00:34:26,280 --> 00:34:28,680 Speaker 1: the innovation, not just to serve the product, but we 632 00:34:28,719 --> 00:34:30,120 Speaker 1: can go and just do what we were doing in 633 00:34:30,160 --> 00:34:33,040 Speaker 1: partnership ourselves. We have the people, we have the compute, 634 00:34:33,080 --> 00:34:37,440 Speaker 1: we have the data, we have everything. Mister Nadella, that 635 00:34:37,560 --> 00:34:39,640 Speaker 1: is an insane thing to say, and the kind of 636 00:34:39,640 --> 00:34:42,600 Speaker 1: thing you'd only say to someone like Kara Swisher. Swisher 637 00:34:42,600 --> 00:34:46,120 Speaker 1: should have made that the headline that Microsoft effectively owns 638 00:34:46,200 --> 00:34:49,319 Speaker 1: open ai. I miss this at the time I found 639 00:34:49,320 --> 00:34:51,640 Speaker 1: out about this quote, like a week or two ago. 640 00:34:52,160 --> 00:34:54,839 Speaker 1: It's crazy to me. It's crazy to me that this 641 00:34:55,239 --> 00:34:58,280 Speaker 1: was said, and it's not front page news. Microsoft owns 642 00:34:58,320 --> 00:35:01,400 Speaker 1: open ai. Who gives a shit about the nonprofit structure. 643 00:35:01,760 --> 00:35:04,480 Speaker 1: But it actually gets better because he followed up that 644 00:35:04,600 --> 00:35:08,680 Speaker 1: quote with this. Microsoft's investment of thirteen billion dollars in 645 00:35:08,760 --> 00:35:12,520 Speaker 1: open ai gives us significant rights, as I said, and 646 00:35:12,600 --> 00:35:15,200 Speaker 1: also this thing, it's not hands off, right. We are 647 00:35:15,200 --> 00:35:17,359 Speaker 1: in there, we are below them, we are above them, 648 00:35:17,440 --> 00:35:20,640 Speaker 1: we are around them. We do the kernel optimizations. Very 649 00:35:20,640 --> 00:35:22,919 Speaker 1: funny if you consider the CrowdStrike thing. By the way 650 00:35:23,480 --> 00:35:26,279 Speaker 1: we build tools, we build the infrastructure. So that's why 651 00:35:26,280 --> 00:35:28,600 Speaker 1: I think a lot of the industrial analysts are saying, 652 00:35:28,680 --> 00:35:31,120 Speaker 1: oh wow, it's really a joint project between Microsoft and 653 00:35:31,120 --> 00:35:34,000 Speaker 1: open ai. The reality is we are as I said, 654 00:35:34,120 --> 00:35:37,759 Speaker 1: very self sufficient in all of this. Well, I'm not 655 00:35:37,840 --> 00:35:41,120 Speaker 1: saying it's for certain. Microsoft can very clearly if they 656 00:35:41,120 --> 00:35:43,799 Speaker 1: so please, simply drop open AI and keep all of 657 00:35:43,800 --> 00:35:48,320 Speaker 1: their stuff. Nevertheless, as the market's sour and jennit of AI, 658 00:35:48,400 --> 00:35:50,720 Speaker 1: I think it's time for big tech to prove something. 659 00:35:51,280 --> 00:35:53,319 Speaker 1: It's time for them to prove that this wasn't all 660 00:35:53,360 --> 00:35:56,440 Speaker 1: a huge, big, stupid, fucking waste of time and money 661 00:35:57,320 --> 00:36:00,120 Speaker 1: to continue with their current paths. Summed up is sh 662 00:36:00,280 --> 00:36:02,600 Speaker 1: of Google and sach and Adella of Microsoft. As the 663 00:36:02,640 --> 00:36:05,600 Speaker 1: loudest of the publicly traded generative AI hype men, will 664 00:36:05,600 --> 00:36:09,520 Speaker 1: have to show everybody something remarkable. But even then they 665 00:36:09,560 --> 00:36:12,719 Speaker 1: will have to show how it actually makes money, and 666 00:36:12,760 --> 00:36:14,680 Speaker 1: if they can't, at least in the case of sach 667 00:36:14,719 --> 00:36:17,680 Speaker 1: and Adella, they can still walk away. As well as 668 00:36:18,520 --> 00:36:20,719 Speaker 1: in the best case scenario, these companies would either have 669 00:36:20,760 --> 00:36:24,200 Speaker 1: to dramatically reduce their costs, something they've really shown they're 670 00:36:24,239 --> 00:36:25,960 Speaker 1: not capable of doing and in fact have said they'll 671 00:36:26,000 --> 00:36:28,520 Speaker 1: do the opposite of or show ways to make generative 672 00:36:28,560 --> 00:36:32,960 Speaker 1: AI so significantly efficient that it somehow finally balances itself out. 673 00:36:33,880 --> 00:36:36,960 Speaker 1: What we're really waiting for now is for somebody to blink. 674 00:36:38,239 --> 00:36:40,560 Speaker 1: If none of the major cloud companies change course over 675 00:36:40,600 --> 00:36:43,640 Speaker 1: the next few months, we could potentially see this boom 676 00:36:43,680 --> 00:36:46,120 Speaker 1: continue for another quarter. And just to be clear, they 677 00:36:46,280 --> 00:36:49,160 Speaker 1: usually in lockstep. They love holding up each other's oligopolies 678 00:36:49,200 --> 00:36:53,480 Speaker 1: and monopolies. They love being a weird cartel that sells data. 679 00:36:53,640 --> 00:36:56,560 Speaker 1: To continue, though, they'll have to commit billions of dollars 680 00:36:56,600 --> 00:36:59,359 Speaker 1: in at the very least cloud credits for open ai 681 00:36:59,400 --> 00:37:02,239 Speaker 1: and Anthropic, as neither of those companies can survive without 682 00:37:02,280 --> 00:37:06,000 Speaker 1: further cash infusions. They'll also have to find some way 683 00:37:06,080 --> 00:37:09,120 Speaker 1: to make any of this useful and profitable, two challenges 684 00:37:09,160 --> 00:37:12,040 Speaker 1: the generative AI has never been able to deal with, 685 00:37:12,400 --> 00:37:16,359 Speaker 1: not even once. And I just as an aside, I've 686 00:37:16,360 --> 00:37:18,399 Speaker 1: had people email and saying, well, I use it for this. 687 00:37:18,680 --> 00:37:20,160 Speaker 1: I was on a show the other days that he 688 00:37:20,239 --> 00:37:23,120 Speaker 1: uses it for show notes. These are all fine whatever, 689 00:37:23,200 --> 00:37:26,400 Speaker 1: These are like little use cases. These people need this 690 00:37:26,440 --> 00:37:29,080 Speaker 1: to make a trillion dollars. Show notes aren't going to 691 00:37:29,080 --> 00:37:32,120 Speaker 1: do it. Shitty translations are not going to do it. 692 00:37:32,760 --> 00:37:35,160 Speaker 1: A picture of Garfield with a gun is not going 693 00:37:35,200 --> 00:37:37,640 Speaker 1: to do it. And big tech is going to have 694 00:37:37,680 --> 00:37:39,640 Speaker 1: to figure out a way to support these companies in 695 00:37:39,640 --> 00:37:42,719 Speaker 1: the face of rising disquiet among shareholders, who in the 696 00:37:42,760 --> 00:37:46,399 Speaker 1: most recent route of AI companies lost them about two 697 00:37:46,440 --> 00:37:50,320 Speaker 1: point six trillion dollars in market capitalization. As I pointed 698 00:37:50,360 --> 00:37:53,240 Speaker 1: out in the Shareholders Supremacy, investors don't really care about 699 00:37:53,239 --> 00:37:56,799 Speaker 1: the future or innovation. They just want profits and their 700 00:37:56,840 --> 00:37:59,319 Speaker 1: portfolios to get bigger. And they want to see the 701 00:37:59,400 --> 00:38:02,800 Speaker 1: number for profit go larger and the costs go lower, 702 00:38:03,000 --> 00:38:07,399 Speaker 1: and absolutely nothing else matters at all, which is why 703 00:38:07,440 --> 00:38:10,040 Speaker 1: they love generative AI at first, because if this actually 704 00:38:10,040 --> 00:38:12,000 Speaker 1: made a bunch of money, this is a way that 705 00:38:12,040 --> 00:38:15,640 Speaker 1: Microsoft could just route cloud costs from themselves to themselves. 706 00:38:16,480 --> 00:38:19,080 Speaker 1: I'll get to that in a future episode about monopolies, though, 707 00:38:20,320 --> 00:38:22,680 Speaker 1: And I think that if generative AI becomes an albatross 708 00:38:22,719 --> 00:38:25,160 Speaker 1: around the next of Google, Microsoft, and Amazon, which I 709 00:38:25,160 --> 00:38:27,520 Speaker 1: really believe it will, these companies are going to face 710 00:38:27,640 --> 00:38:31,040 Speaker 1: really nasty pressure to curtail their investments and their capex. 711 00:38:31,800 --> 00:38:34,400 Speaker 1: Investors will ask why these companies are spending a collective 712 00:38:34,640 --> 00:38:37,480 Speaker 1: one hundred two hundred billion dollars by the end of 713 00:38:37,480 --> 00:38:40,320 Speaker 1: twenty twenty five on a technology that likely won't drive 714 00:38:40,400 --> 00:38:42,600 Speaker 1: much revenue growth when it could be used for things 715 00:38:42,640 --> 00:38:46,160 Speaker 1: like stock dividends and stock buybacks, and I don't know 716 00:38:46,880 --> 00:38:49,840 Speaker 1: thrones for the demon lords that live underneath the offices. 717 00:38:49,920 --> 00:38:53,080 Speaker 1: I don't know how Microsoft truly makes money. That's a joke, 718 00:38:53,280 --> 00:38:56,320 Speaker 1: don't email me about it. But I believe in any case, 719 00:38:56,800 --> 00:38:58,560 Speaker 1: the next month is going to be critical to the 720 00:38:58,560 --> 00:39:01,560 Speaker 1: future of the AI boom, and I believe this is 721 00:39:01,560 --> 00:39:05,320 Speaker 1: a time for industry wide introspection and to really consider 722 00:39:05,440 --> 00:39:08,920 Speaker 1: everyone's roles and why this bubble existed in the first place. 723 00:39:09,920 --> 00:39:11,840 Speaker 1: In the next episode, I'm going to take a closer 724 00:39:11,880 --> 00:39:14,799 Speaker 1: look into open Aiy's likely demise, what it looks like, 725 00:39:15,200 --> 00:39:17,880 Speaker 1: and why the company's view avenues for salvation in fact 726 00:39:18,080 --> 00:39:21,080 Speaker 1: look like dead ends. And I'll get into how fucking 727 00:39:21,239 --> 00:39:25,040 Speaker 1: ridiculous this whole bubble is and what its potential collapse 728 00:39:25,080 --> 00:39:35,120 Speaker 1: spells for the future of the tech industry. Thank you 729 00:39:35,160 --> 00:39:37,880 Speaker 1: for listening to Better Offline. The editor and composer of 730 00:39:37,880 --> 00:39:40,879 Speaker 1: the Better Offline theme song is Matasowski. You can check 731 00:39:40,880 --> 00:39:43,840 Speaker 1: out more of his music and audio projects at Matasowski 732 00:39:43,880 --> 00:39:47,359 Speaker 1: dot com, M A T T O. S O. W 733 00:39:47,719 --> 00:39:51,200 Speaker 1: Ski dot com. You can email me at easy at 734 00:39:51,280 --> 00:39:53,920 Speaker 1: Better Offline dot com or visit Better Offline dot com 735 00:39:53,920 --> 00:39:56,400 Speaker 1: to find more podcast links and of course my newsletter. 736 00:39:56,760 --> 00:39:59,239 Speaker 1: I also really recommend you go to chat dot Where's 737 00:39:59,280 --> 00:40:01,759 Speaker 1: youoeed dot com to visit the discord, and go to 738 00:40:01,800 --> 00:40:05,160 Speaker 1: our slash Better Offline to check out I'll Reddit. Thank 739 00:40:05,239 --> 00:40:08,640 Speaker 1: you so much for listening. Better Offline is a production 740 00:40:08,719 --> 00:40:11,360 Speaker 1: of cool Zone Media. For more from cool Zone Media, 741 00:40:11,480 --> 00:40:14,839 Speaker 1: visit our website Coolzonemedia dot com, or check us out 742 00:40:14,840 --> 00:40:17,839 Speaker 1: on the iHeartRadio app, Apple Podcasts, or wherever you get 743 00:40:17,880 --> 00:40:18,560 Speaker 1: your podcasts.