1 00:00:01,840 --> 00:00:04,360 Speaker 1: Corey Doctor Rowe is the author and tech critic who 2 00:00:04,360 --> 00:00:06,200 Speaker 1: brought us the term and shitification. 3 00:00:06,920 --> 00:00:09,719 Speaker 2: First, a company is good to its end users, then 4 00:00:10,080 --> 00:00:12,440 Speaker 2: it locks them in, and then it is good to 5 00:00:12,480 --> 00:00:15,760 Speaker 2: its business customers, who then get locked in, and then 6 00:00:15,920 --> 00:00:18,759 Speaker 2: it takes all the value for itself, leaving behind a 7 00:00:18,800 --> 00:00:22,320 Speaker 2: giant pile of shit that is in shitification, a tragedy 8 00:00:22,360 --> 00:00:23,400 Speaker 2: in three acts. 9 00:00:24,000 --> 00:00:26,120 Speaker 1: Now his warning that what we've been led to believe 10 00:00:26,239 --> 00:00:30,400 Speaker 1: is the inevitable march of AI towards world domination. Isn't 11 00:00:30,400 --> 00:00:34,320 Speaker 1: that inevitable after all? And actually what we're marching towards 12 00:00:34,520 --> 00:00:39,360 Speaker 1: could be the end of the AI bubble Because every 13 00:00:39,440 --> 00:00:42,559 Speaker 1: time you ask Chat, GPT, Claude or Gemini a question, 14 00:00:43,080 --> 00:00:45,479 Speaker 1: companies like open ai are losing money. 15 00:00:45,760 --> 00:00:47,840 Speaker 3: The more we use, the more they lose. 16 00:00:48,440 --> 00:00:51,440 Speaker 1: And while those AI giants burn through billions of dollars 17 00:00:51,479 --> 00:00:54,400 Speaker 1: a year gambling that eventually they'll figure out how to 18 00:00:54,400 --> 00:00:57,800 Speaker 1: turn it into a profitable business, the world's economy is 19 00:00:57,840 --> 00:01:04,600 Speaker 1: hanging in the balance. I'm Ruby Jones, and you're listening 20 00:01:04,680 --> 00:01:08,959 Speaker 1: to seven AM today Corey doctor Oh on why the 21 00:01:08,959 --> 00:01:11,320 Speaker 1: AI boom is being pushed into every part of our 22 00:01:11,360 --> 00:01:21,000 Speaker 1: lives and what happens when the bubble bursts. It's Tuesday, 23 00:01:21,319 --> 00:01:26,399 Speaker 1: July twenty eighth. 24 00:01:27,080 --> 00:01:29,640 Speaker 3: Corey, Welcome to seven AM. Thanks for coming on the show. 25 00:01:29,800 --> 00:01:31,280 Speaker 2: Oh well, thank you for having me on. 26 00:01:31,640 --> 00:01:34,120 Speaker 1: So I thought we could begin by talking about some 27 00:01:34,200 --> 00:01:39,360 Speaker 1: of the claims that are made around AI and its inevitability. 28 00:01:40,040 --> 00:01:42,480 Speaker 1: The way it's sort of framed is this technology that 29 00:01:42,520 --> 00:01:43,120 Speaker 1: has arrived. 30 00:01:43,160 --> 00:01:45,240 Speaker 3: It's going to change the nature of work. 31 00:01:45,280 --> 00:01:48,320 Speaker 1: It's going to fundamentally reshape life. 32 00:01:48,120 --> 00:01:48,760 Speaker 3: As we know it. 33 00:01:48,920 --> 00:01:51,040 Speaker 1: So to begin with, tell me a bit about who's 34 00:01:51,040 --> 00:01:52,840 Speaker 1: making these claims and why they are. 35 00:01:53,400 --> 00:01:57,280 Speaker 2: Well, I think that we've heard these claims before. We 36 00:01:57,360 --> 00:01:59,720 Speaker 2: heard that cryptocurrency was going to replace all of our 37 00:01:59,760 --> 00:02:02,560 Speaker 2: trains actions, that the metaverse was going to replace the 38 00:02:02,560 --> 00:02:06,960 Speaker 2: web that NFTs, where we're going to replace all art. 39 00:02:07,080 --> 00:02:12,239 Speaker 4: We're talking about AI physics or AI biology for drug discovery, 40 00:02:12,560 --> 00:02:17,320 Speaker 4: AI agents for customer service and support of diagnosis, and 41 00:02:17,400 --> 00:02:20,120 Speaker 4: of course physical AI robotic systems. 42 00:02:20,440 --> 00:02:23,800 Speaker 2: This kind of claim to declare a fatal complete even 43 00:02:23,840 --> 00:02:27,920 Speaker 2: before you've got started, is very powerful rhetorically. All of 44 00:02:27,960 --> 00:02:31,840 Speaker 2: these AI capabilities that make you smarter, help you communicate better, 45 00:02:32,280 --> 00:02:35,440 Speaker 2: improve your memory, improve your senses, and more. It's a 46 00:02:35,440 --> 00:02:39,440 Speaker 2: way to exterminate any thought that things might be different 47 00:02:39,600 --> 00:02:42,120 Speaker 2: before you can even start thinking about it. 48 00:02:42,200 --> 00:02:46,200 Speaker 5: Anthropic, we talk a lot about the exponential, and I 49 00:02:46,200 --> 00:02:49,480 Speaker 5: think that's what we're all feeling right now. The jumps 50 00:02:49,560 --> 00:02:53,239 Speaker 5: keep getting bigger and the intervals keep getting shorter. 51 00:02:53,240 --> 00:02:56,000 Speaker 2: And we hear it now from AI people who are 52 00:02:56,040 --> 00:03:00,160 Speaker 2: desperately trying to explain how they can take a sector 53 00:03:00,200 --> 00:03:04,359 Speaker 2: that grosses fifty billion American dollars a year and somehow 54 00:03:04,400 --> 00:03:06,880 Speaker 2: turn it into a profit making venture, despite the fact 55 00:03:06,919 --> 00:03:09,520 Speaker 2: that they've spent more than a trillion dollars just this year, 56 00:03:10,120 --> 00:03:12,400 Speaker 2: and despite the fact that every generation of it is 57 00:03:12,480 --> 00:03:16,040 Speaker 2: less profitable than the previous generation, and despite the fact 58 00:03:16,120 --> 00:03:18,480 Speaker 2: that they're having to replace all of their hardware every 59 00:03:18,520 --> 00:03:19,320 Speaker 2: two to three years. 60 00:03:19,400 --> 00:03:22,360 Speaker 6: Sard from opening I CEO Sam Altman with a fresh 61 00:03:22,400 --> 00:03:25,040 Speaker 6: look at the company's financials. He says that they expect 62 00:03:25,040 --> 00:03:28,600 Speaker 6: their annualized revenue run rate to top twenty billion dollars 63 00:03:28,600 --> 00:03:31,680 Speaker 6: this year it was thirteen billion dollars before that, and 64 00:03:31,760 --> 00:03:34,600 Speaker 6: to grow to hundreds of billions by twenty thirty. That's 65 00:03:34,639 --> 00:03:35,920 Speaker 6: also a revision upward. 66 00:03:36,040 --> 00:03:38,360 Speaker 1: Yeah, so this kind of talk, it's led to what 67 00:03:38,720 --> 00:03:42,480 Speaker 1: many people have speculated is in AI bubble, but believing 68 00:03:42,920 --> 00:03:45,520 Speaker 1: big tech when it talks about what AI can and 69 00:03:45,640 --> 00:03:48,120 Speaker 1: will do. We've now arrived at this place where we 70 00:03:48,200 --> 00:03:52,119 Speaker 1: have vastly overestimated its value. So what is your view 71 00:03:52,120 --> 00:03:55,800 Speaker 1: on whether AI is actually providing what it's promising the 72 00:03:55,800 --> 00:03:57,280 Speaker 1: productivity and profit. 73 00:03:57,560 --> 00:03:59,960 Speaker 2: Well, the productivity and profit are nowhere to be found 74 00:04:00,240 --> 00:04:04,880 Speaker 2: in the world. You have some ANNEC data from individuals 75 00:04:04,920 --> 00:04:07,520 Speaker 2: who say, well, I've used AI and made myself far 76 00:04:07,560 --> 00:04:11,520 Speaker 2: more productive. But when you actually look at the firms 77 00:04:11,560 --> 00:04:14,640 Speaker 2: that have deployed AI, what you find is that even 78 00:04:14,680 --> 00:04:17,720 Speaker 2: with the massive subsidy the AI companies are offering, and 79 00:04:17,800 --> 00:04:21,520 Speaker 2: the subsidies can't be overstated. Basically, at this point, AI 80 00:04:21,560 --> 00:04:24,279 Speaker 2: companies are selling hundred dollar bills at a dollar apiece, 81 00:04:24,880 --> 00:04:29,240 Speaker 2: and even with that, you're not finding the productivity gains 82 00:04:29,279 --> 00:04:32,279 Speaker 2: that firms were hoping for. We keep seeing firms that 83 00:04:32,400 --> 00:04:35,080 Speaker 2: fire large numbers of workers and then have to try 84 00:04:35,080 --> 00:04:38,880 Speaker 2: and hire them back. But even worse was about a 85 00:04:38,960 --> 00:04:41,680 Speaker 2: month ago when the AI companies were all thinking about 86 00:04:41,680 --> 00:04:44,520 Speaker 2: their IPOs and they said, well, let's clean up our 87 00:04:44,560 --> 00:04:46,840 Speaker 2: balance sheets a little and raise the price of these 88 00:04:46,960 --> 00:04:50,120 Speaker 2: hundred dollar bills to five dollars apiece, and all of 89 00:04:50,160 --> 00:04:52,960 Speaker 2: the CEOs, even the ones who've been most enthusiastic about 90 00:04:53,080 --> 00:04:55,800 Speaker 2: using AI, suddenly turned around and said, wait, your one 91 00:04:55,839 --> 00:04:58,520 Speaker 2: hundred dollar bills were barely worth it at a dollar each, 92 00:04:58,600 --> 00:05:02,800 Speaker 2: but at five dollars each, there's just no way we're 93 00:05:02,839 --> 00:05:04,719 Speaker 2: going to be able to continue to buy these things. 94 00:05:04,760 --> 00:05:06,800 Speaker 2: And they suddenly went from saying you're going to get 95 00:05:06,800 --> 00:05:09,240 Speaker 2: fired if you don't use AI to you're going to 96 00:05:09,279 --> 00:05:11,440 Speaker 2: get fired if you keep using AI, at least in 97 00:05:11,440 --> 00:05:12,920 Speaker 2: the way you've been using it so far. 98 00:05:13,200 --> 00:05:16,400 Speaker 1: Yeah, so you're saying these companies AI companies like open 99 00:05:16,440 --> 00:05:20,440 Speaker 1: Ai and Anthropic, they aren't actually making real profits. In fact, 100 00:05:20,800 --> 00:05:24,039 Speaker 1: they're losing billions of dollars a year, yet the valuation 101 00:05:24,160 --> 00:05:27,800 Speaker 1: of their stock is skyrocketing. So what is the impact 102 00:05:27,880 --> 00:05:28,080 Speaker 1: of that. 103 00:05:28,839 --> 00:05:32,039 Speaker 2: Yeah, Well, the shares and affirm that are growing are 104 00:05:32,040 --> 00:05:35,600 Speaker 2: worth far more than the shares of affirm that is mature. 105 00:05:36,160 --> 00:05:37,919 Speaker 2: And that's because a share is a claim on the 106 00:05:37,960 --> 00:05:40,359 Speaker 2: future income of a company. So if you can imagine 107 00:05:40,360 --> 00:05:42,920 Speaker 2: two companies, each of which turns over a million dollars 108 00:05:42,960 --> 00:05:45,560 Speaker 2: this year, but one of which is expected to double 109 00:05:45,600 --> 00:05:48,279 Speaker 2: in size next year, you could see why shares in 110 00:05:48,320 --> 00:05:50,960 Speaker 2: that company that's doubling in size would be worth more 111 00:05:51,000 --> 00:05:54,800 Speaker 2: than the shares in the company that are static. And 112 00:05:55,200 --> 00:05:58,320 Speaker 2: when your shares enjoy what's called this high price to 113 00:05:58,400 --> 00:06:02,240 Speaker 2: earnings ratio, when they're valued this highly, they become extremely 114 00:06:02,360 --> 00:06:05,120 Speaker 2: liquid and you can use them in place of cash. 115 00:06:05,400 --> 00:06:08,360 Speaker 2: And the advantage of shares over dollars is that shares 116 00:06:08,400 --> 00:06:11,119 Speaker 2: are an endogenous product. They're made within the four walls 117 00:06:11,120 --> 00:06:13,880 Speaker 2: of the company through the simple expedient of typing zeros 118 00:06:13,920 --> 00:06:17,400 Speaker 2: into spreadsheets. And the corollary of this is that when 119 00:06:17,400 --> 00:06:20,800 Speaker 2: you stop growing, when you become mature, which has to 120 00:06:20,839 --> 00:06:23,640 Speaker 2: happen eventually. Right, if you're Google with a ninety percent 121 00:06:23,680 --> 00:06:25,799 Speaker 2: search market share, you're not going to grow your search 122 00:06:25,839 --> 00:06:29,440 Speaker 2: market share anymore, right, And so you know, when you 123 00:06:29,560 --> 00:06:33,320 Speaker 2: cap out your growth, your company becomes grossly overvalued because 124 00:06:33,320 --> 00:06:35,920 Speaker 2: it's been valued as a growth company, but now it's 125 00:06:35,960 --> 00:06:38,200 Speaker 2: a mature company. And at that point you see these 126 00:06:38,200 --> 00:06:41,040 Speaker 2: panic sell offs. This is why every time tech companies 127 00:06:41,320 --> 00:06:45,400 Speaker 2: announce very small slowdowns in their growth or reversals, you 128 00:06:45,520 --> 00:06:48,000 Speaker 2: get these mass selloffs. You know, Facebook had a two 129 00:06:48,040 --> 00:06:50,919 Speaker 2: hundred and forty billion dollars one day sell off in 130 00:06:50,960 --> 00:06:53,200 Speaker 2: the first quarter of twenty twenty two, after announcing it 131 00:06:53,240 --> 00:06:56,039 Speaker 2: had slightly fewer American signups than they'd anticipate. 132 00:06:56,080 --> 00:06:58,839 Speaker 5: It is on Wall Street this morning after shares of 133 00:06:58,880 --> 00:07:01,320 Speaker 5: Facebook's parent Come company plunged yesterday. 134 00:07:01,520 --> 00:07:04,279 Speaker 2: Is stock continuing to tumble this week after a disappointing 135 00:07:04,320 --> 00:07:05,680 Speaker 2: report this quarter, shares. 136 00:07:05,480 --> 00:07:09,679 Speaker 1: In Facebook now called Meta, plunging twenty six percent after 137 00:07:09,760 --> 00:07:13,320 Speaker 1: results showed people are turning away from the social media platform. 138 00:07:13,400 --> 00:07:15,800 Speaker 2: It was the largest decapitalization of any firm in the 139 00:07:15,840 --> 00:07:18,560 Speaker 2: history of markets until it happened again to Nvidia a 140 00:07:18,600 --> 00:07:20,840 Speaker 2: couple of years later, this time to the tune of 141 00:07:21,080 --> 00:07:24,600 Speaker 2: three quarters of a trillion dollars. But if you can't grow, well, 142 00:07:24,600 --> 00:07:26,400 Speaker 2: the next best thing is to promise you're going to 143 00:07:26,400 --> 00:07:29,480 Speaker 2: grow right, to say, well, look we're about to conquer 144 00:07:29,560 --> 00:07:32,680 Speaker 2: a new imaginary market in NFTs or web three, or 145 00:07:32,960 --> 00:07:36,120 Speaker 2: now that AI is getting a little stale as super intelligence. 146 00:07:36,200 --> 00:07:38,640 Speaker 2: And you know, all of these things benefit in part 147 00:07:38,680 --> 00:07:42,080 Speaker 2: from having technical meanings that aren't well understood or even 148 00:07:42,080 --> 00:07:44,480 Speaker 2: well defined, So you can claim that you're going to 149 00:07:44,480 --> 00:07:47,160 Speaker 2: conquer a market that no one else understands, which means 150 00:07:47,160 --> 00:07:49,040 Speaker 2: when they say that that market isn't worth what you 151 00:07:49,080 --> 00:07:52,400 Speaker 2: say it is. You can repost by saying, nuh uh. 152 00:07:52,440 --> 00:07:54,080 Speaker 2: I'm the only person in the world who even knows 153 00:07:54,080 --> 00:07:55,920 Speaker 2: what this market is, so how can you tell me 154 00:07:55,960 --> 00:07:57,640 Speaker 2: it's not as big as I claim it's going to be. 155 00:07:58,320 --> 00:08:01,160 Speaker 1: But the scale here when we're talking about AI is 156 00:08:01,200 --> 00:08:03,280 Speaker 1: where this gets a bit frightening. I mean, you say, 157 00:08:03,280 --> 00:08:06,520 Speaker 1: we've reached the point where the US economy, which really, 158 00:08:06,520 --> 00:08:09,840 Speaker 1: by extension means the world economy, depends on AI and 159 00:08:09,880 --> 00:08:10,560 Speaker 1: its growth. 160 00:08:10,720 --> 00:08:13,240 Speaker 3: So how precarious is that? 161 00:08:13,680 --> 00:08:16,800 Speaker 2: Oh, it's extremely precarious. This is the frightening thing about AI. 162 00:08:16,880 --> 00:08:18,760 Speaker 2: It's not like that we're going to teach too many 163 00:08:18,800 --> 00:08:20,840 Speaker 2: words to the word guessing program and it's going to 164 00:08:20,840 --> 00:08:23,160 Speaker 2: wake up and become God and turn us into paper 165 00:08:23,160 --> 00:08:27,160 Speaker 2: clips or something. It's that seven firms, six of whom 166 00:08:27,200 --> 00:08:30,800 Speaker 2: are losing money to the seventh one represent thirty five 167 00:08:30,840 --> 00:08:32,640 Speaker 2: percent of the S and P five hundred, the most 168 00:08:32,679 --> 00:08:36,560 Speaker 2: important financial index in the world. They are all passing 169 00:08:36,559 --> 00:08:39,040 Speaker 2: around the same one hundred billion dollar IOU and pretending 170 00:08:39,080 --> 00:08:40,800 Speaker 2: that it's in all of their bank accounts at once, 171 00:08:40,840 --> 00:08:43,080 Speaker 2: which you know, to call that accounting fraud is to 172 00:08:43,120 --> 00:08:46,240 Speaker 2: do violence to the noble accounting fraud. And at the 173 00:08:46,280 --> 00:08:49,760 Speaker 2: same time, every new generation of AI is less profitable 174 00:08:49,760 --> 00:08:52,640 Speaker 2: than the previous one. Every new customer for AI costs 175 00:08:52,679 --> 00:08:56,440 Speaker 2: THEII companies money, and every time the AI companies have 176 00:08:56,520 --> 00:09:00,400 Speaker 2: a customer that uses their products, they lose even more money. Meanwhile, 177 00:09:00,440 --> 00:09:03,880 Speaker 2: the actual hard assets, the data centers, the GPUs, they're 178 00:09:03,960 --> 00:09:06,520 Speaker 2: being replaced every like two to three years, and I 179 00:09:06,600 --> 00:09:10,199 Speaker 2: think we underestimate just how much of that hard capital 180 00:09:10,240 --> 00:09:13,880 Speaker 2: has to be thrown away with every generation of the technology. 181 00:09:14,240 --> 00:09:16,760 Speaker 2: And often it's the case that to retrofit a data 182 00:09:16,800 --> 00:09:19,240 Speaker 2: center it would cost more than scraping into the foundation 183 00:09:19,360 --> 00:09:21,200 Speaker 2: slab and building a new one when you buy new 184 00:09:21,280 --> 00:09:25,199 Speaker 2: chips for it. So all of this stuff is basically disposable, 185 00:09:25,480 --> 00:09:27,560 Speaker 2: and you add it all up and it just doesn't 186 00:09:27,640 --> 00:09:30,360 Speaker 2: turn into a business you could bail out. It's not 187 00:09:30,400 --> 00:09:32,840 Speaker 2: like buying a bunch of distressed real estate assets that 188 00:09:32,880 --> 00:09:36,000 Speaker 2: when you're done, are real estate that people can live in. 189 00:09:36,559 --> 00:09:39,440 Speaker 2: It's like buying a toxic waste dump that's on fire 190 00:09:39,559 --> 00:09:41,199 Speaker 2: and you don't know how to put out the fire. 191 00:09:41,200 --> 00:09:43,480 Speaker 2: In every year, you have to commit more resources to 192 00:09:43,960 --> 00:09:47,600 Speaker 2: evacuating people from the region because it's poisoning their soil 193 00:09:47,640 --> 00:09:48,520 Speaker 2: and their air. 194 00:09:50,960 --> 00:09:51,480 Speaker 3: Coming up. 195 00:09:51,600 --> 00:09:53,880 Speaker 1: What will happen to workers in the AI boom and 196 00:09:54,080 --> 00:10:04,880 Speaker 1: eventual bust. So Corey, you in your book make the 197 00:10:04,920 --> 00:10:08,600 Speaker 1: case that nothing is inevitable about AI, and the best 198 00:10:08,600 --> 00:10:10,800 Speaker 1: way to think about it is in terms of whether 199 00:10:11,160 --> 00:10:13,160 Speaker 1: we are driving it to work for us or whether 200 00:10:13,200 --> 00:10:16,800 Speaker 1: it's being used to force us to change our behavior. 201 00:10:17,120 --> 00:10:18,360 Speaker 3: Talk me through that idea. 202 00:10:18,640 --> 00:10:20,560 Speaker 2: Yeah, for sure. You know, this isn't the first time 203 00:10:20,720 --> 00:10:23,600 Speaker 2: labor and automation have come into conflict, and one of 204 00:10:23,640 --> 00:10:27,560 Speaker 2: its bedrocks is that when labor drives automation adoption, typically 205 00:10:27,559 --> 00:10:30,160 Speaker 2: it's in service to making things better. Workers want to 206 00:10:30,200 --> 00:10:34,679 Speaker 2: improve their outputs. When capital drives automation, they want to 207 00:10:34,720 --> 00:10:38,560 Speaker 2: maximize quantity, not quality, and often it's quantity at the 208 00:10:38,600 --> 00:10:43,080 Speaker 2: expense of quality. And so we see automation now in 209 00:10:43,160 --> 00:10:46,600 Speaker 2: AI being driven almost exclusively by capital. You know, when 210 00:10:46,600 --> 00:10:48,959 Speaker 2: you read the business press, it's not like the early 211 00:10:49,000 --> 00:10:52,040 Speaker 2: two thousands, where you know, Harvard Business Review was publishing 212 00:10:52,320 --> 00:10:55,000 Speaker 2: panicked articles about what CEOs we're going to do about 213 00:10:55,000 --> 00:10:57,160 Speaker 2: all the young workers who expected to use the web 214 00:10:57,160 --> 00:10:59,679 Speaker 2: at the office. Now it's full of articles about how 215 00:10:59,760 --> 00:11:03,160 Speaker 2: c are threatening to fire workers unless they use AI. Right, 216 00:11:03,559 --> 00:11:06,760 Speaker 2: So you know this is very much capital driven automation. 217 00:11:07,400 --> 00:11:10,840 Speaker 2: But when you are the weakest link in the chain, 218 00:11:10,880 --> 00:11:13,920 Speaker 2: when you're the bottleneck and the output of the machine, 219 00:11:14,360 --> 00:11:16,280 Speaker 2: then you're going to be worked at the absolute limit 220 00:11:16,320 --> 00:11:19,240 Speaker 2: of your capacity by a machine that by definition can 221 00:11:19,320 --> 00:11:21,360 Speaker 2: last longer than you, move faster than you, and is 222 00:11:21,400 --> 00:11:23,520 Speaker 2: stronger than you, which means that the minute you make 223 00:11:23,559 --> 00:11:26,240 Speaker 2: a mistake, you're in for a world of hurt. You know, 224 00:11:26,360 --> 00:11:29,480 Speaker 2: Amazon built the world's most automated warehouses. They're also the 225 00:11:29,480 --> 00:11:33,240 Speaker 2: warehouses with the highest rate of injury. That's not a coincidence. 226 00:11:33,320 --> 00:11:36,319 Speaker 2: It's the consequence you don't pay eight figures for warehouse 227 00:11:36,320 --> 00:11:39,000 Speaker 2: automation and run the machine slower than you need to. 228 00:11:39,280 --> 00:11:41,440 Speaker 2: And when you run the machine at the maximum speed 229 00:11:41,440 --> 00:11:43,920 Speaker 2: of the people who operate in it and around it, 230 00:11:44,240 --> 00:11:46,760 Speaker 2: it means that the minutes someone gets something wrong, they 231 00:11:46,760 --> 00:11:48,160 Speaker 2: get impaled on a forklift. 232 00:11:48,960 --> 00:11:51,520 Speaker 1: Can we talk a little more about workers, because your 233 00:11:51,520 --> 00:11:55,720 Speaker 1: contention here seems to be that tech companies are pushing 234 00:11:55,880 --> 00:11:58,679 Speaker 1: the exponential growth of AI because growth is key to 235 00:11:58,720 --> 00:12:02,240 Speaker 1: their survival. To do that, they are in turn convincing 236 00:12:02,360 --> 00:12:05,600 Speaker 1: bosses of companies that AI can do the jobs of 237 00:12:05,679 --> 00:12:08,360 Speaker 1: their workers, meaning that they'll be able to either fire 238 00:12:08,400 --> 00:12:11,200 Speaker 1: those workers or pay them less, leading to more profit 239 00:12:11,240 --> 00:12:14,280 Speaker 1: for the company. So where does all of this leave workers. 240 00:12:14,800 --> 00:12:16,840 Speaker 2: Yeah, well, you know, I think you just have to 241 00:12:16,920 --> 00:12:19,840 Speaker 2: understand that, like, bosses are infinitely horny to get rid 242 00:12:19,880 --> 00:12:23,079 Speaker 2: of workers because every day they have ego shattering confrontations 243 00:12:23,320 --> 00:12:25,200 Speaker 2: with people who know how to do things that they 244 00:12:25,200 --> 00:12:27,640 Speaker 2: don't know how to do, and who you know, tell 245 00:12:27,679 --> 00:12:29,520 Speaker 2: them every time they show up with a great idea 246 00:12:29,600 --> 00:12:31,839 Speaker 2: that it's illegal or moral or it's going to kill 247 00:12:31,880 --> 00:12:34,880 Speaker 2: a bunch of people. And you can you know, dispel 248 00:12:35,040 --> 00:12:37,000 Speaker 2: that fear that you're in the back seat with a 249 00:12:37,080 --> 00:12:39,880 Speaker 2: Fisher Price steering wheel instead of in the driver's seat 250 00:12:40,160 --> 00:12:42,880 Speaker 2: by using AI to like come up with the product 251 00:12:42,880 --> 00:12:45,040 Speaker 2: idea and the AI craps it out and you don't 252 00:12:45,040 --> 00:12:47,840 Speaker 2: have to like actually talk to people about it. And 253 00:12:47,880 --> 00:12:50,840 Speaker 2: what you get is a very dim world for workers. 254 00:12:50,880 --> 00:12:53,240 Speaker 2: Which is not to say that AI isn't useful for workers. 255 00:12:53,600 --> 00:12:56,120 Speaker 2: I mean, I think the way to resolve the paradox 256 00:12:56,320 --> 00:12:59,480 Speaker 2: of for example, programmers who sometimes say, well, I use 257 00:12:59,520 --> 00:13:01,640 Speaker 2: AI and my code is better than it's ever been. 258 00:13:01,720 --> 00:13:03,960 Speaker 2: I can't believe how much I'm getting done, and other 259 00:13:04,040 --> 00:13:07,200 Speaker 2: programmers who are also skilled and also reliable narrators of 260 00:13:07,240 --> 00:13:09,800 Speaker 2: their experience, who say, I can't believe how much tech 261 00:13:09,840 --> 00:13:12,240 Speaker 2: debt we're incurring at my company. You know, we make 262 00:13:12,520 --> 00:13:16,040 Speaker 2: whatever av onics never get on an airplane again, because 263 00:13:16,080 --> 00:13:19,400 Speaker 2: we have shoveled so much bad code into the world's 264 00:13:19,720 --> 00:13:22,800 Speaker 2: civilian aircraft that they should never be trusted to fly. 265 00:13:23,280 --> 00:13:24,720 Speaker 2: And you know, you look at those two groups of 266 00:13:24,720 --> 00:13:28,920 Speaker 2: people and you know, try to resolve this seeming paradox 267 00:13:28,960 --> 00:13:31,720 Speaker 2: in what you find inevitably is that the first group 268 00:13:31,720 --> 00:13:33,880 Speaker 2: of workers are choosing how they use AI, and the 269 00:13:33,880 --> 00:13:37,120 Speaker 2: second group of workers, well, they've seen all their colleagues fired, 270 00:13:37,160 --> 00:13:38,960 Speaker 2: and they're being asked to work at ten times the 271 00:13:39,000 --> 00:13:41,840 Speaker 2: former speed, to mark the AI's homework and take the 272 00:13:41,880 --> 00:13:45,000 Speaker 2: blame when it goes wrong. And so, you know, I 273 00:13:45,040 --> 00:13:47,720 Speaker 2: think that there's plenty of useful things we can do 274 00:13:47,800 --> 00:13:49,480 Speaker 2: with AI. I just don't think we're going to do 275 00:13:49,559 --> 00:13:52,040 Speaker 2: them in the conditions in which the imperative is to 276 00:13:52,120 --> 00:13:54,040 Speaker 2: fire workers and replace them with software. 277 00:13:54,679 --> 00:13:57,240 Speaker 1: So we're at the point now where the tech itself 278 00:13:57,280 --> 00:14:00,800 Speaker 1: is being used by corporations. Sometimes it works expense but 279 00:14:00,880 --> 00:14:05,880 Speaker 1: it's likely vastly over valued. So what happens if that 280 00:14:05,920 --> 00:14:06,800 Speaker 1: bubble bursts? 281 00:14:07,320 --> 00:14:10,120 Speaker 2: Well, you know, not all bubbles are created equally. If 282 00:14:10,160 --> 00:14:13,520 Speaker 2: you remember the Enron bubble, where you had this energy 283 00:14:13,679 --> 00:14:16,840 Speaker 2: trading firm that was just doing accounting fraud and sucking 284 00:14:16,880 --> 00:14:19,320 Speaker 2: up the money of ordinary people who were just trying 285 00:14:19,320 --> 00:14:21,240 Speaker 2: to save so that they wouldn't starve to death or 286 00:14:21,280 --> 00:14:24,480 Speaker 2: be homeless when they got old. When they popped, there 287 00:14:24,480 --> 00:14:27,800 Speaker 2: was nothing left behind for all the billions they stole. Meanwhile, 288 00:14:27,840 --> 00:14:30,000 Speaker 2: there was another fraud at the same time called WorldCom, 289 00:14:30,480 --> 00:14:32,800 Speaker 2: and WorldCom is also an accounting fraud, but wasn't a 290 00:14:32,800 --> 00:14:36,040 Speaker 2: pure accounting fraud. WorldCom's claim was they had hundreds of 291 00:14:36,040 --> 00:14:39,280 Speaker 2: billions of dollars in fiber optic orders, and to kind 292 00:14:39,280 --> 00:14:42,640 Speaker 2: of make the show look good, they dug up streets 293 00:14:42,640 --> 00:14:45,040 Speaker 2: all over the world and put fiber optic in the ground. 294 00:14:45,520 --> 00:14:47,960 Speaker 2: And that fiber optic is still there. So when you 295 00:14:48,000 --> 00:14:51,600 Speaker 2: look at contemporary bubbles like cryptocurrency, it's not really going 296 00:14:51,640 --> 00:14:54,440 Speaker 2: to leave anything behind. When crypto goes to zero, we'll 297 00:14:54,480 --> 00:14:58,280 Speaker 2: have like ugly monkey JPEGs and stupid Austrian economics. But 298 00:14:58,400 --> 00:15:00,880 Speaker 2: when AI goes to zero, was still have data centers, 299 00:15:00,880 --> 00:15:03,080 Speaker 2: we'll still have GPUs, We'll have a lot of workers 300 00:15:03,120 --> 00:15:05,800 Speaker 2: who know how to do stuff, and we're going to 301 00:15:05,800 --> 00:15:07,920 Speaker 2: have these open source models that have barely been touched, 302 00:15:08,000 --> 00:15:10,720 Speaker 2: barely been optimized, and every time someone looks at them, 303 00:15:11,240 --> 00:15:13,040 Speaker 2: they can find all kinds of ways to make them 304 00:15:13,080 --> 00:15:16,200 Speaker 2: perform better. And I think, if anything, we'll probably see 305 00:15:16,240 --> 00:15:19,040 Speaker 2: some pretty good stuff as a result of that. Notwithstanding 306 00:15:19,040 --> 00:15:21,320 Speaker 2: that the economic crisis is going to be terrible, and 307 00:15:21,640 --> 00:15:23,960 Speaker 2: when a third of the stock market is vaporized because 308 00:15:24,000 --> 00:15:26,600 Speaker 2: these firms, you know, have to finally mark their assets 309 00:15:26,600 --> 00:15:29,560 Speaker 2: to market and stop pretending that their IOUs are and 310 00:15:29,560 --> 00:15:32,280 Speaker 2: all their bank accounts at once, we'll probably see governments 311 00:15:32,280 --> 00:15:34,680 Speaker 2: around the world do austerity again, which is just going 312 00:15:34,720 --> 00:15:37,080 Speaker 2: to drive more people into the arms of fascists. But 313 00:15:37,120 --> 00:15:39,200 Speaker 2: in terms of the technology, it'll probably be a bit 314 00:15:39,280 --> 00:15:42,320 Speaker 2: like Web two point zero, where all the stupid bosses 315 00:15:42,440 --> 00:15:44,720 Speaker 2: left the Bay Area and you could buy the servers 316 00:15:44,720 --> 00:15:47,160 Speaker 2: they left behind for pennies on the dollar, and all 317 00:15:47,160 --> 00:15:49,640 Speaker 2: the workers who had interesting ideas for what to do 318 00:15:49,960 --> 00:15:52,520 Speaker 2: suddenly didn't have to convince their bosses that these were 319 00:15:52,520 --> 00:15:54,880 Speaker 2: good ideas, and you know, they hired each other and 320 00:15:54,880 --> 00:15:57,280 Speaker 2: made some interesting stuff. You know. I was living in 321 00:15:57,280 --> 00:16:00,200 Speaker 2: San Francisco when the dot com bubble burst. One of 322 00:16:00,240 --> 00:16:02,840 Speaker 2: the things that I'd always craved during that whole bubble 323 00:16:03,000 --> 00:16:06,480 Speaker 2: was these very fancy ergonomic chairs from Steelcase called the 324 00:16:06,560 --> 00:16:09,480 Speaker 2: leap Chair, that were about fifteen hundred dollars each. And 325 00:16:09,520 --> 00:16:12,080 Speaker 2: I remember a day walking through the mission where I lived, 326 00:16:12,120 --> 00:16:14,000 Speaker 2: and there was a dot com ceo in front of 327 00:16:14,040 --> 00:16:17,280 Speaker 2: the office where he just lost his lease, selling like 328 00:16:17,360 --> 00:16:20,160 Speaker 2: a fleet of Steelcase leap chairs still in the plastic 329 00:16:20,200 --> 00:16:22,920 Speaker 2: for twenty five dollars apiece. I bought six of them 330 00:16:22,920 --> 00:16:24,680 Speaker 2: and use them as a dining room set for the 331 00:16:24,720 --> 00:16:27,240 Speaker 2: next ten years. Right, you know the idea of like 332 00:16:27,400 --> 00:16:30,640 Speaker 2: buying AI now, it's like buying your Steelcase leaf chairs 333 00:16:30,680 --> 00:16:33,240 Speaker 2: in like the spring of twenty twenty, when you can 334 00:16:33,280 --> 00:16:35,560 Speaker 2: see the fall of twenty twenty coming at you like 335 00:16:35,600 --> 00:16:37,680 Speaker 2: a freight train, when you're gonna be able to buy 336 00:16:37,680 --> 00:16:41,240 Speaker 2: things for pennies. Right, just just wait. If you really 337 00:16:41,280 --> 00:16:44,320 Speaker 2: feel you need a sovereign AI, just wait and buy 338 00:16:44,360 --> 00:16:45,400 Speaker 2: it then. 339 00:16:46,000 --> 00:16:48,200 Speaker 3: Well, Corey, thank you so much for speaking with me. 340 00:16:48,480 --> 00:16:49,640 Speaker 2: Well, thank you very much. 341 00:16:52,080 --> 00:16:52,320 Speaker 3: Cory. 342 00:16:52,400 --> 00:16:55,440 Speaker 1: Doctor Rowe is in Australia. This August for appearances at 343 00:16:55,480 --> 00:16:58,760 Speaker 1: Sydney's Festival of Dangerous Ideas and in Melbourne at the 344 00:16:58,760 --> 00:17:01,240 Speaker 1: Capitol presented by the Wheel Center and now or ever, 345 00:17:01,760 --> 00:17:04,480 Speaker 1: his new book, The Reverse Centaur's Guide to Life After 346 00:17:04,520 --> 00:17:20,280 Speaker 1: AI is out now. Also when the news Victoria's Premier 347 00:17:20,320 --> 00:17:22,920 Speaker 1: to Center Allen could today be rolled just months out 348 00:17:22,960 --> 00:17:25,800 Speaker 1: from the November state election. A group of ministers and 349 00:17:25,920 --> 00:17:28,960 Speaker 1: MPs met with the Premier yesterday urging her to stand down. 350 00:17:29,520 --> 00:17:32,240 Speaker 1: Deputy Premier Ben Carroll has advised the Premier that he 351 00:17:32,280 --> 00:17:34,480 Speaker 1: intends to be a candidate for leadership in the event 352 00:17:34,520 --> 00:17:37,080 Speaker 1: of a spill, which is likely to unfold a caucus 353 00:17:37,080 --> 00:17:41,240 Speaker 1: meeting today and South Australia has recorded seven new suspected 354 00:17:41,280 --> 00:17:44,639 Speaker 1: cases of H five bird flu. If further testing of 355 00:17:44,680 --> 00:17:48,000 Speaker 1: the cases by the CSIRO comes back positive, it'll take 356 00:17:48,040 --> 00:17:50,919 Speaker 1: the total number of bird flu detections in South Australia 357 00:17:51,000 --> 00:17:54,240 Speaker 1: to fourteen. Today, there have been no H five bird 358 00:17:54,280 --> 00:17:57,800 Speaker 1: flu detections in the poultry industry, captive birds, pet or 359 00:17:57,840 --> 00:18:02,400 Speaker 1: other wildlife and livestock andmory industries and wildlife organizations say 360 00:18:02,440 --> 00:18:06,600 Speaker 1: they're working alongside the state government to state vigilant I'm 361 00:18:06,680 --> 00:18:07,320 Speaker 1: Ruby Jones. 362 00:18:07,400 --> 00:18:09,159 Speaker 3: This is seven AM. Thanks for listening.