WEBVTT - Managing The Situation 0:00:02.340 --> 0:00:07.880 Hello and welcome to Better Offline, I'm your host Ed Zitron. 0:00:15.910 --> 0:00:17.889 This week is monologue week, by which I mean I 0:00:17.930 --> 0:00:19.790 had a guest pull out last minute and so you 0:00:19.810 --> 0:00:22.990 have to hear me talk about stuff. As ever, subscribe 0:00:23.010 --> 0:00:25.550 to the Premium Newsletter. There's a discount link in the notes. 0:00:25.570 --> 0:00:27.730 It means a lot to me. It's my main source 0:00:27.770 --> 0:00:30.130 of income, along with the podcast. And like and subscribe 0:00:30.150 --> 0:00:32.169 if you're on YouTube. Or toss me a glert on 0:00:32.330 --> 0:00:36.900 Sloob if you're on Sloob. I'm also on Ploik and Dwork. Anyway, 0:00:37.320 --> 0:00:39.519 today I'm going to talk about a problem in basically 0:00:39.580 --> 0:00:43.380 every part of the world. Managing situations versus actually dealing 0:00:43.420 --> 0:00:47.250 with the complexities of the world itself. A couple weeks ago, 0:00:47.290 --> 0:00:50.470 I wrote a piece called Hyperscale Normalization, inspired, as you 0:00:50.510 --> 0:00:54.300 may have guessed, by the Adam Curtis documentary Hypernormalization, where 0:00:54.310 --> 0:00:57.380 he describes, and I quote, a world where, over the 0:00:57.420 --> 0:01:02.640 past 40 years, politicians, financiers, and technological utopians, rather than 0:01:02.700 --> 0:01:05.600 face up to the real complexities of the world, retreated, 0:01:06.220 --> 0:01:09.069 constructing what he calls a simpler version of the world 0:01:09.370 --> 0:01:11.670 in order to hang on to power, with the world 0:01:11.709 --> 0:01:15.860 going along with it because the simplicity was reassuring. It's 0:01:15.900 --> 0:01:17.920 a two and a half hour long documentary, but it 0:01:17.959 --> 0:01:20.320 comes down to a simple point. That we as a 0:01:20.380 --> 0:01:23.699 society manage problems rather than dealing with the messy and 0:01:23.720 --> 0:01:26.380 complex elements of the world around us. Putting off problems 0:01:26.440 --> 0:01:29.200 again and again and under the assumption that they'll disappear 0:01:29.240 --> 0:01:31.750 if we do so enough times. We use the system 0:01:31.810 --> 0:01:35.669 to manage these problems rather than actually eradicating them or 0:01:36.010 --> 0:01:39.809 dealing with the conditions. And of course, they never actually disappear. 0:01:39.870 --> 0:01:44.070 The problems actually compound. Let's take the subprime mortgage crisis 0:01:44.090 --> 0:01:49.870 for example. While post-2008 financial regulations, Dodd-Frank and the now-weakened 0:01:49.930 --> 0:01:53.360 Volcker Rule theoretically helped curb some of the excesses that 0:01:53.420 --> 0:01:56.560 led to the Great Financial Crisis, most of the people 0:01:56.640 --> 0:02:01.840 involved remained both wealthy and employed, blissfully unblackballed from the 0:02:01.900 --> 0:02:04.480 world of finance or indeed by the media. In fact, 0:02:05.240 --> 0:02:07.600 I can find nobody of that time in the media 0:02:07.660 --> 0:02:11.270 who was ousted or even really critiqued for how badly 0:02:11.310 --> 0:02:16.310 they fucked up. While financial institutions and companies saw over 0:02:16.310 --> 0:02:21.110 $ 14 trillion in bailouts, only $ 46 billion, about 10% of 0:02:21.150 --> 0:02:25.140 the TARP plan, focused on trying to save homeowners from foreclosure. 0:02:25.240 --> 0:02:28.820 And in the end, to quote a congressional panel from 2009, 0:02:28.820 --> 0:02:31.340 there was no evidence that the Treasury has used TARP 0:02:31.400 --> 0:02:35.639 funds to support the housing market by avoiding preventable foreclosures. 0:02:37.090 --> 0:02:39.929 Government funds were used to manage a systemic crisis by 0:02:39.990 --> 0:02:42.880 making sure the system survived, rather than building a better 0:02:42.910 --> 0:02:47.560 society or fixing the system or really doing anything more 0:02:47.600 --> 0:02:51.760 than the Dodd-Frank and Volcker-style stuff that really only slightly 0:02:51.820 --> 0:02:54.180 gnawed on the problems rather than chewing them up and 0:02:54.220 --> 0:02:57.660 eating them. Because doing so, feeding the system money, cleaning 0:02:57.720 --> 0:03:01.060 up its messes, explaining away its successes, is easier than 0:03:01.139 --> 0:03:03.799 having any kind of vision or ideal to aspire to. 0:03:05.120 --> 0:03:08.860 Doing so strips out, to quote Curtis, the intractable complexities 0:03:08.919 --> 0:03:12.040 of the real world by creating a simpler one. One 0:03:12.080 --> 0:03:15.780 where the system works, that progress always flows through following 0:03:15.820 --> 0:03:19.549 the course of, often poorly remembered, history and backing whatever 0:03:19.590 --> 0:03:23.829 the system trumpets that moment. Everything ultimately returns to a 0:03:23.889 --> 0:03:26.790 point where someone will say the system has worked this 0:03:26.850 --> 0:03:30.360 long and things always work out in the end, even 0:03:30.400 --> 0:03:33.769 if they really haven't done so, because thinking of an alternative, 0:03:34.070 --> 0:03:36.050 even as you see proof point after proof point that 0:03:36.070 --> 0:03:39.930 the system is not functioning right and keeps having crises created, 0:03:40.490 --> 0:03:42.840 is near impossible due to the prominence of the dogma 0:03:42.880 --> 0:03:46.660 and confidence from those in power. The great financial crisis 0:03:46.700 --> 0:03:49.800 was managed by dropping interest rates to zero and bailing 0:03:49.860 --> 0:03:54.480 out banks and automakers and basically any systemic companies that 0:03:54.520 --> 0:03:58.210 would pose a risk to America or various societies, rather 0:03:58.250 --> 0:04:02.050 than the people in the societies. COVID was managed by 0:04:02.090 --> 0:04:05.970 the loosest lockdown protocols possible, giving businesses hundreds of billions 0:04:06.010 --> 0:04:09.070 of dollars via the Paycheck Protection Program. And just to 0:04:09.110 --> 0:04:13.450 be clear, only for salaried employees rather than contractors. giving 0:04:13.490 --> 0:04:16.690 banks and financial institutions trillions of dollars via the very 0:04:16.750 --> 0:04:19.290 same systems that were used in the great financial crisis, 0:04:19.650 --> 0:04:24.180 and basically doing absolutely nothing to fix the actual problems 0:04:24.630 --> 0:04:31.500 that COVID kind of exposed. The wealth inequality, racial injustice, discrimination, 0:04:31.560 --> 0:04:33.700 these things mostly were just kind of pushed off to 0:04:33.720 --> 0:04:37.060 the side. Companies picked up DEI as long as necessary 0:04:37.320 --> 0:04:40.110 to appease markets that believed there was actual social change 0:04:40.150 --> 0:04:43.570 without actually ingesting it. At no point was the system 0:04:43.610 --> 0:04:47.090 itself significantly altered, such as creating Medicare for All or 0:04:47.110 --> 0:04:50.430 a social safety net or making price gouging illegal, which 0:04:50.470 --> 0:04:52.950 meant that the problems didn't really go away so much 0:04:53.010 --> 0:04:56.719 as they were managed. The net result was that in 2021, 0:04:56.720 --> 0:05:00.240 the financial system got saturated with easy money, trillions of 0:05:00.260 --> 0:05:02.440 dollars of capital that made the rich richer and the 0:05:02.480 --> 0:05:06.940 poor poorer. Meanwhile, corporations have been using supply chain crises 0:05:07.000 --> 0:05:10.289 and the specter of inflation to raise prices across the board, 0:05:10.370 --> 0:05:12.910 though it became obvious that the real reason was sheer, 0:05:13.270 --> 0:05:17.270 unabated greed, with companies posting record profits to soak up 0:05:17.310 --> 0:05:20.210 the cash from a frothy society excited to be back 0:05:20.250 --> 0:05:23.530 in the real world. A real world that had materially 0:05:23.630 --> 0:05:26.230 changed since the beginning of COVID, but no one really 0:05:26.270 --> 0:05:29.370 wanted to acknowledge that. By the end of 2022, the 0:05:29.529 --> 0:05:33.820 Federal Reserve would raise interest rates by a dramatic 4.25%, 0:05:33.820 --> 0:05:36.260 leading to tens of thousands of people losing their jobs 0:05:36.320 --> 0:05:40.150 in the tech sector alone. And things have never really recovered. 0:05:40.990 --> 0:05:43.610 Regular people have spent years watching the price of goods 0:05:43.690 --> 0:05:47.849 increase due to inflation, even after those interest rate bumps, 0:05:47.910 --> 0:05:50.469 despite the fact that the increase in pricing was mostly 0:05:50.510 --> 0:05:55.120 driven by, get this, corporations raising prices. Yet some parts 0:05:55.180 --> 0:05:57.779 of the legacy media spent an alarming amount of time 0:05:57.900 --> 0:06:00.940 chiding their readers for thinking otherwise, even going against their 0:06:00.980 --> 0:06:04.450 own reporting as a means of providing balanced coverage. insisting 0:06:04.529 --> 0:06:07.480 again and again that the economy is good, contorting to 0:06:07.520 --> 0:06:10.600 prove that prices aren't actually higher, even as the companies 0:06:10.640 --> 0:06:15.820 like Pepsi boasted about literally raising their prices. In fact, 0:06:15.839 --> 0:06:19.010 the media spent years debating it with itself, whether price 0:06:19.050 --> 0:06:21.790 gouging was happening, despite years of proof that it was 0:06:21.890 --> 0:06:25.469 and actually continues to this day. The system, by which 0:06:25.529 --> 0:06:28.029 I mean the conjoined forces of the media, the markets, 0:06:28.070 --> 0:06:31.670 and the American government, focused on aggressively forcing everything back 0:06:31.730 --> 0:06:33.429 to where it used to be, or at least making 0:06:33.470 --> 0:06:36.250 it look that way. When things were rough, it pumped 0:06:36.300 --> 0:06:39.180 money into the system. When things seemed too frothy, it 0:06:39.200 --> 0:06:41.580 made that money harder to come by, which ended up 0:06:42.000 --> 0:06:46.460 punishing regular people for the largest part. For the most part, 0:06:46.940 --> 0:06:49.620 regular people in general were punished for the excesses of 0:06:49.660 --> 0:06:52.830 the system itself. When they took advantage of remote work, 0:06:52.890 --> 0:06:55.940 job flexibility, and purchasing power, they were told they were lazy, 0:06:56.020 --> 0:06:58.440 that they were wrong, that this was temporary, and that 0:06:58.500 --> 0:07:00.820 in fact they were too greedy with what they were 0:07:00.860 --> 0:07:06.350 generously given, and indeed expected too much from society. The 0:07:06.390 --> 0:07:10.830 continued perpetuation of the system is always right ultimately led 0:07:10.910 --> 0:07:15.430 to the media near completely detaching from reality. A regular 0:07:15.470 --> 0:07:19.850 person experiencing aggressively worsening standards in their own lives would 0:07:19.930 --> 0:07:22.470 open the news only to be sneered at for feeling bad, 0:07:22.590 --> 0:07:25.270 mocked for questioning the numbers, told to sit down and 0:07:25.370 --> 0:07:27.910 shut up because the media and those that informed them 0:07:28.190 --> 0:07:31.850 knew better. After two years of hope and abundance, the 0:07:31.890 --> 0:07:35.330 world and the system itself attempted to revert everyone back 0:07:35.370 --> 0:07:38.370 to the before times, all without the veneer of prosperity 0:07:38.470 --> 0:07:43.420 or progress or, indeed, never going back. In other words, 0:07:43.760 --> 0:07:46.520 a regular person's form of reality was a confusing mess 0:07:46.560 --> 0:07:50.810 of social media, alternative media, mainstream media, and governments that 0:07:50.870 --> 0:07:53.610 seemed intent on saying that either everything was fine or 0:07:53.650 --> 0:07:56.330 that there was a specific thing to blame, usually a 0:07:56.710 --> 0:08:00.250 foreigner or some sort of listener or reader themselves, or 0:08:00.290 --> 0:08:04.750 really just people expecting too much. No critiques are levied 0:08:04.790 --> 0:08:07.580 at the system itself, or the choices made by it, 0:08:07.830 --> 0:08:10.050 or the way in which the media chooses to cover 0:08:10.090 --> 0:08:14.020 systemic movements, because to do so is considered either stupid, 0:08:14.180 --> 0:08:16.840 because the system always works right, or hopeless, because the 0:08:16.880 --> 0:08:33.650 system is just so powerful. So what does this have 0:08:33.690 --> 0:08:37.800 to do with AI exactly? Okay, so I want you 0:08:37.820 --> 0:08:41.460 to think about how people defend the AI industry for 0:08:41.500 --> 0:08:44.340 a second. When you do that, it's all going to 0:08:44.360 --> 0:08:47.850 begin to make sense. If you talk about how OpenAI 0:08:47.890 --> 0:08:50.510 and Anthropic are losing tens of billions of dollars a year, 0:08:50.820 --> 0:08:53.760 you'll get told that Uber and Amazon Web Services lost 0:08:53.800 --> 0:08:57.660 a lot of money. Even though both of those services burned, what, 0:08:57.660 --> 0:09:01.120 $ 30 billion each? And in Amazon Web Services' case, it 0:09:01.200 --> 0:09:04.780 was $ 29. 7 billion of total capex between the year it 0:09:05.040 --> 0:09:09.100 started in 2003 and the year it became profitable in 2015. 0:09:09.100 --> 0:09:13.429 Doesn't matter, though. Doesn't matter. The system's always right. If 0:09:13.480 --> 0:09:17.679 you say that generative AI's hallucinations make massive mistakes that 0:09:17.740 --> 0:09:21.320 make LLMs inherently untrustworthy, everyone will tell you it's the 0:09:21.360 --> 0:09:24.969 early days. If you say you're not finding use cases, 0:09:25.350 --> 0:09:28.670 you're castigated. You're told actually you need to use it 0:09:28.790 --> 0:09:32.600 more so that you can find it useful. This artificial 0:09:32.679 --> 0:09:36.800 intelligence is actually kind of stupid, but it's also super smart. 0:09:37.100 --> 0:09:39.469 You see, it's smart enough to do stuff if you 0:09:39.500 --> 0:09:41.880 make it do it, but not smart enough to actually 0:09:41.920 --> 0:09:44.490 help you in the way that we associate with software, 0:09:44.610 --> 0:09:49.670 or indeed, the memetic form of artificial intelligence. If you 0:09:49.690 --> 0:09:53.810 critique basically anything about generative AI, you're given a shot 0:09:53.850 --> 0:09:57.209 of systemic apologia. It's early. It's just like the early 0:09:57.250 --> 0:09:59.670 days of the internet. These are the smartest people in 0:09:59.710 --> 0:10:02.920 the world. These are the richest and largest companies in 0:10:02.960 --> 0:10:05.760 the world. And otherwise you need to shut the fuck 0:10:05.860 --> 0:10:09.440 up and trust the process because that's what's always worked. 0:10:10.660 --> 0:10:14.480 Even those who are anti-AI fall into the same systemic traps. 0:10:15.450 --> 0:10:18.790 The amount of times I hear from people suggesting that 0:10:18.830 --> 0:10:22.490 this is a massive multi-trillion dollar data center build-up that's 0:10:22.710 --> 0:10:27.130 actually a big secret government surveillance program. Well, those people 0:10:27.170 --> 0:10:30.670 are falling into the very same systemic thinking. Believing that 0:10:30.740 --> 0:10:33.619 because a bunch of money is being spent on illusory 0:10:33.660 --> 0:10:37.560 demand and circular financing, that in fact, the system isn't stupid. 0:10:37.840 --> 0:10:40.980 That these companies aren't dumb. That actually there's a big plan. 0:10:41.280 --> 0:10:44.760 The smart people running the system have a big secret plan. 0:10:45.790 --> 0:10:48.349 They don't. Just want to be clear that they don't. 0:10:48.830 --> 0:10:52.310 But this is, again, convenient systemic thinking. Because when you 0:10:52.350 --> 0:10:55.190 don't think in this way, well, the real world seems 0:10:55.250 --> 0:10:59.800 kind of insane. Yet every systemic failure has come from 0:10:59.860 --> 0:11:03.760 a lack of any real plan, strategy, or forethought. The 0:11:03.800 --> 0:11:07.420 great financial crisis happened because everybody assumed that the system 0:11:07.660 --> 0:11:11.360 had everything covered. That the financial institutions had hedged their 0:11:11.380 --> 0:11:15.520 bets correctly. That everybody made sure that these were high-quality investments. 0:11:15.880 --> 0:11:18.500 that anyone with an adjustable-rate mortgage would be able to 0:11:18.540 --> 0:11:23.520 refinance it or sell their ever-appreciating house, mantra after mantra 0:11:23.559 --> 0:11:26.140 based on a lack of curiosity and complete faith in 0:11:26.179 --> 0:11:28.990 the various systems that would eventually fail them, as they 0:11:29.010 --> 0:11:33.550 had failed before. We are watching the very same thing 0:11:33.670 --> 0:11:38.550 happen again with AI. Yesterday, the Financial Times reported that 0:11:38.610 --> 0:11:42.890 bankers for OpenAI and Anthropix IPOs were seeking investment-grade ratings 0:11:42.929 --> 0:11:47.260 for both companies, pushing agencies to rate two unprofitable, unsustainable 0:11:47.330 --> 0:11:51.809 companies as if they were, well, real companies. Per the FT, 0:11:52.050 --> 0:11:54.690 the argument is that the two companies' public listings would 0:11:54.750 --> 0:11:57.750 unlock vast amounts of liquidity and improve the health of 0:11:57.790 --> 0:12:00.960 their balance sheets, by which they mean, by giving them 0:12:01.010 --> 0:12:03.860 access to lots of debt, both OpenAI and Anthropic would 0:12:03.920 --> 0:12:07.280 be able to take on lots of debt. That's really it. 0:12:08.679 --> 0:12:11.960 This is a deeply dangerous way of managing two companies 0:12:12.000 --> 0:12:15.800 that will eventually run out of money. Anthropic and OpenAI 0:12:15.820 --> 0:12:17.840 will need hundreds of billions of dollars a year to 0:12:17.880 --> 0:12:22.480 meet their $ 1. 3 trillion of compute commitments. And even if 0:12:22.520 --> 0:12:26.020 they're somehow rated investment grade, and by the way, S & P, Moody's, 0:12:26.090 --> 0:12:29.430 if you're listening, if you fucking do this, you deserve 0:12:29.510 --> 0:12:32.950 nothing but suffering. If any of you have any chance 0:12:32.990 --> 0:12:35.410 you hear this, fuck you guys. You're going to do 0:12:35.429 --> 0:12:39.050 it again, aren't you? But let's say they become investment grade. 0:12:40.110 --> 0:12:43.970 They're going to need so much debt. So much debt 0:12:44.070 --> 0:12:47.970 with onerously high interest, because there'll be one wrong above junk. 0:12:48.450 --> 0:12:50.990 Which means that unless they magically find ways to become 0:12:51.050 --> 0:12:55.510 profitable and generate massive amounts of cash, I don't see 0:12:55.570 --> 0:12:57.670 how they service the debt they're going to need. And 0:12:57.720 --> 0:13:01.679 I must be clear, their compute commitments, OpenAI is expected 0:13:01.720 --> 0:13:06.540 to spend $ 750 billion through 2030. How do they afford that? 0:13:07.500 --> 0:13:10.260 They lose tens of billions of dollars a year. I'm 0:13:10.280 --> 0:13:14.459 considered a firebrand for asking these questions. And these companies, 0:13:14.520 --> 0:13:18.740 like I said, have over $ 1. 3 trillion in compute commitments, 0:13:19.100 --> 0:13:21.680 which means, like I said, they're going to own hundreds 0:13:22.000 --> 0:13:27.070 of billions of dollars of compute payments by 2030. An 0:13:27.110 --> 0:13:30.990 unbelievable sum that dwarfs the operating expenses of Meta at 0:13:31.150 --> 0:13:35.450 around $ 141 billion a year, and Microsoft at about $ 175 0:13:35.150 --> 0:13:39.120 billion a year. In fact, based on the projections, I'm 0:13:39.179 --> 0:13:41.800 pretty sure Anthropica OpenAI is going to spend more than 0:13:41.820 --> 0:13:44.770 those two numbers combined at At some point. And when 0:13:44.820 --> 0:13:47.400 I say spend, intend to spend. They're going to need 0:13:47.420 --> 0:13:51.060 the money, which they don't have. Both companies are seeing 0:13:51.140 --> 0:13:54.460 growth decelerate and 80% of their enterprise revenues per ramp, 0:13:54.500 --> 0:13:57.160 as I discussed last week, come from the top 1% 0:13:57.160 --> 0:14:00.640 of their customers who are predominantly unprofitable AI startups that 0:14:00.660 --> 0:14:03.780 are subsidizing their users' AI token spend. Meaning that for 0:14:03.840 --> 0:14:06.600 every dollar a user spends in subscriptions on these services, 0:14:06.620 --> 0:14:09.320 they're able to burn several dollars in AI tokens. with 0:14:09.340 --> 0:14:12.360 all that money flowing back to Anthropic and OpenAI, who 0:14:12.480 --> 0:14:16.520 are also supported by venture capital funding. The same venture 0:14:16.540 --> 0:14:22.960 capital funding that their customers need. Hmm. Any attempts to 0:14:22.980 --> 0:14:26.480 mark OpenAI or Anthropic as investment-grade will simply delay the 0:14:26.540 --> 0:14:29.500 inevitable point at which they run out of money. This 0:14:29.540 --> 0:14:31.460 wouldn't have been the case if these AI labs were 0:14:31.480 --> 0:14:34.100 much smaller companies, or had much smaller losses, or, to 0:14:34.120 --> 0:14:39.610 be specific, hadn't signed $ 1. 3 trillion of commitments. Said commitments 0:14:39.640 --> 0:14:42.630 are also take-or-pay agreements that mean that regardless of whether 0:14:42.650 --> 0:14:45.450 they have the demand to actually need the data centers, 0:14:45.850 --> 0:14:50.070 Anthropic and OpenAI will have to pay Amazon, Google, Microsoft, Cerebrus, CoreWeave, 0:14:50.290 --> 0:14:52.850 or any number of other providers once the capacity comes online. 0:14:54.210 --> 0:14:57.270 If they can afford it, that is. However you may 0:14:57.330 --> 0:14:59.810 feel about me or the greater AI bubble is immaterial 0:14:59.850 --> 0:15:02.000 to the fact that everything will seem like it's fine 0:15:02.080 --> 0:15:04.260 right up until somebody can't raise money and make a 0:15:04.280 --> 0:15:07.580 payment to either a Neo, Cloud, Hyperscaler, or AI lab. 0:15:08.770 --> 0:15:11.690 For this to keep working, AI startups must continue to 0:15:11.730 --> 0:15:13.970 be able to raise hundreds of billions of dollars every 0:15:14.010 --> 0:15:17.170 few months to make single-digit or tens of billions of 0:15:17.210 --> 0:15:20.610 dollars a month, all as Anthropic and OpenAI must also 0:15:20.670 --> 0:15:22.910 continue to raise tens or hundreds of billions of dollars 0:15:23.190 --> 0:15:26.210 to pay hyperscalers for compute so that they can, in 0:15:26.250 --> 0:15:28.570 addition to raising hundreds of billions of dollars of debt, 0:15:28.890 --> 0:15:32.270 spend that money on GPUs from NVIDIA and TPUs from Broadcom, 0:15:32.650 --> 0:15:34.820 who can only continue to make hundreds of billions of 0:15:34.880 --> 0:15:37.040 dollars a year in revenue as long as They can 0:15:37.100 --> 0:15:40.180 either provide justifications for lenders to keep issuing those hundreds 0:15:40.220 --> 0:15:42.820 of billions of dollars in debt, or backstop the data 0:15:42.880 --> 0:15:46.180 sent as the debt will get spent on. In other words, 0:15:46.240 --> 0:15:48.240 the AI bubble is based on the whims of maybe 0:15:48.280 --> 0:15:51.230 a few hundred companies spending money on two companies to 0:15:51.260 --> 0:15:54.550 justify five companies spending money with one company. Or two 0:15:54.590 --> 0:15:56.790 if you count Broadcom, which you don't have to if 0:15:56.810 --> 0:15:59.390 you don't want to. I don't like thinking about Broadcom 0:15:59.450 --> 0:16:02.860 or Hoctan very much. Anyone who tells you not to 0:16:02.900 --> 0:16:06.130 worry about Anthropic, a company that loses billions of dollars 0:16:06.190 --> 0:16:08.170 a year, if not tens of billions of dollars a year, 0:16:08.210 --> 0:16:12.490 and has made $ 517 billion in commute commitments, is a 0:16:12.530 --> 0:16:16.250 con artist. And anyone who prints a quote saying that 0:16:16.670 --> 0:16:18.690 you shouldn't have to think about it too hard, or 0:16:18.710 --> 0:16:21.010 you shouldn't have to worry about it, or that this 0:16:21.070 --> 0:16:24.500 is just like the dot-com bubble, really doesn't give a 0:16:24.580 --> 0:16:27.680 shit about whether you live or die. But that really 0:16:27.720 --> 0:16:30.350 is the current state of the tech industry. A death 0:16:30.380 --> 0:16:34.410 cult obsessed with growth, empowered by a media ecosystem obsessed 0:16:34.430 --> 0:16:37.630 with measuring and celebrating how much it's growing and what 0:16:37.670 --> 0:16:40.650 might grow it in the future, always framed in the 0:16:40.710 --> 0:16:44.530 terms set by the rich and powerful. The failure of 0:16:44.670 --> 0:16:47.730 everyone here to meet this moment with clarity and purpose 0:16:48.010 --> 0:16:51.110 will lead to a market correction that dwarfs the dot-com 0:16:51.150 --> 0:16:54.690 bubble easily, exposing many of those involved as a phony, 0:16:54.790 --> 0:16:58.100 a fraud, an imbecile, a ghoul, a coward, or utterly 0:16:58.220 --> 0:17:03.620 impossibly ignorant. Anyway, I'll be back Friday with another monologue. 0:17:04.280 --> 0:17:04.899 I love you all.