1 00:00:02,480 --> 00:00:06,720 Speaker 1: All zo media. You know what time it is. It's 2 00:00:06,720 --> 00:00:08,800 Speaker 1: your better offline monologue for the week, and I'm ed 3 00:00:08,880 --> 00:00:19,680 Speaker 1: z Tron. Now. Today's monologue is an attempt to bring 4 00:00:19,720 --> 00:00:21,920 Speaker 1: you further into my work, which I've kind of already 5 00:00:21,960 --> 00:00:25,120 Speaker 1: been doing with multiple listeners materially contributing to the reporting 6 00:00:25,160 --> 00:00:28,160 Speaker 1: behind this show. So many of you are eager to help, 7 00:00:28,200 --> 00:00:30,319 Speaker 1: and if you want to help, well, reach out to 8 00:00:30,360 --> 00:00:34,040 Speaker 1: me with information and do so on signal ez itron 9 00:00:34,120 --> 00:00:38,879 Speaker 1: dot seventy six. That's ez it r o N or 10 00:00:38,920 --> 00:00:41,159 Speaker 1: e z it r o N if you're Canadian or 11 00:00:41,200 --> 00:00:44,720 Speaker 1: British dot seventy six on signal. Also, before I go 12 00:00:44,800 --> 00:00:46,919 Speaker 1: any further, I've had a lot of emails and reddits 13 00:00:46,960 --> 00:00:49,240 Speaker 1: and as such, if you're wondering about my thoughts on 14 00:00:49,280 --> 00:00:52,560 Speaker 1: the Open Ai Disney deal, they are as followers. Wow, 15 00:00:52,680 --> 00:00:55,640 Speaker 1: Disney invested a billion dollars in open Ai. That's about 16 00:00:55,760 --> 00:00:59,200 Speaker 1: month's worth of inference. This only delays the inevitable. No 17 00:00:59,240 --> 00:01:02,920 Speaker 1: one cares. They're threatening Google with legal action. Google will 18 00:01:02,920 --> 00:01:05,400 Speaker 1: settle and they'll do exactly the same crap with VO. 19 00:01:05,880 --> 00:01:09,080 Speaker 1: Nothing will happen, a bunch of money will get lost. Similarly, 20 00:01:09,080 --> 00:01:11,759 Speaker 1: if you've heard about time naming AI executives as its 21 00:01:11,840 --> 00:01:15,120 Speaker 1: Person of the Year. Remember that Mark Bennioff of Salesforce 22 00:01:15,280 --> 00:01:19,440 Speaker 1: Huji Booster owns the publication and literally ran advertorial for 23 00:01:19,560 --> 00:01:25,520 Speaker 1: Salesforce's agent Force. Wank. It's nothing, nothing's happening. It's just 24 00:01:25,600 --> 00:01:27,800 Speaker 1: more money being passed around, so we can all get 25 00:01:27,840 --> 00:01:31,119 Speaker 1: blown on nothing. But back to the rest of the monologue. 26 00:01:31,480 --> 00:01:33,360 Speaker 1: I'm going to tell you about the things I need 27 00:01:33,400 --> 00:01:36,800 Speaker 1: to do a better job. Not don't literally mean the research. 28 00:01:36,840 --> 00:01:39,320 Speaker 1: I mean, if you know this stuff, I need your help. 29 00:01:39,920 --> 00:01:42,000 Speaker 1: Let's start with a big one. If you have first 30 00:01:42,080 --> 00:01:46,360 Speaker 1: hand knowledge of GPUs being warehouse, this is Nvidia AI GPUs. 31 00:01:46,440 --> 00:01:50,080 Speaker 1: Please God, if you reach out with this about gaming GPUs, 32 00:01:50,120 --> 00:01:52,320 Speaker 1: I'll be genuinely pissed off of you. But I need 33 00:01:52,360 --> 00:01:54,440 Speaker 1: to know about this. I want pictures if you can 34 00:01:54,440 --> 00:01:56,560 Speaker 1: get them. I want numbers. No, I want to know 35 00:01:56,720 --> 00:01:59,400 Speaker 1: whose GPUs they are and ideally what models A one 36 00:01:59,480 --> 00:02:01,680 Speaker 1: hundreds eight one hundred's probably not going to be that, 37 00:02:01,960 --> 00:02:05,560 Speaker 1: it's going to be those blackweld GPUs if you see them. Similarly, 38 00:02:05,640 --> 00:02:09,200 Speaker 1: if you've heard anything about Nvidia's aigpus that might not 39 00:02:09,280 --> 00:02:11,600 Speaker 1: be well known, reach out especially if it pertains to 40 00:02:11,600 --> 00:02:14,280 Speaker 1: how much they cost to run. Similarly, if you know 41 00:02:14,560 --> 00:02:18,040 Speaker 1: anything about Nvidia's blackweld GPUs that you suspect I don't, 42 00:02:18,040 --> 00:02:20,600 Speaker 1: please do reach out. I had somebody recently tell me 43 00:02:20,639 --> 00:02:24,760 Speaker 1: about remarkable failure rates and I want to know more. Similarly, costs. 44 00:02:24,919 --> 00:02:28,440 Speaker 1: I cannot ask for costs enough. Help me. Please, if 45 00:02:28,440 --> 00:02:31,520 Speaker 1: you have any information pretending to the revenues behind open Ai, 46 00:02:31,639 --> 00:02:35,760 Speaker 1: Anthropic or any major AI company, please get in touch. Similarly, 47 00:02:35,840 --> 00:02:37,440 Speaker 1: I want to hear from you if you work at 48 00:02:37,480 --> 00:02:40,800 Speaker 1: any of the large cloud companies and know anything about 49 00:02:40,800 --> 00:02:43,840 Speaker 1: the AI revenues or indeed the associated costs of running 50 00:02:43,840 --> 00:02:47,480 Speaker 1: AI on Google Cloud, Amazon Web Services and Microsoft Asure. 51 00:02:47,960 --> 00:02:50,679 Speaker 1: And yeah, if you have anything from open ai or 52 00:02:50,680 --> 00:02:53,119 Speaker 1: Anthropic about what they might be paying show as your 53 00:02:53,320 --> 00:02:56,440 Speaker 1: Google Cloud or Amazon Web services, please please please reach 54 00:02:56,480 --> 00:02:59,760 Speaker 1: out to me. Similarly, if you have anything of the 55 00:02:59,760 --> 00:03:03,000 Speaker 1: same from Corweave or Lambda or Nebius or any of those, 56 00:03:03,520 --> 00:03:06,800 Speaker 1: please reach out. Also, if you work at an AI startup, 57 00:03:06,840 --> 00:03:09,079 Speaker 1: one that pays open ai or Anthropic to use their 58 00:03:09,120 --> 00:03:11,920 Speaker 1: services or models, please tell me about those costs. We 59 00:03:11,960 --> 00:03:14,440 Speaker 1: don't know very much about them at all, and your 60 00:03:14,520 --> 00:03:17,919 Speaker 1: thoughts and your help, well, it'd be very much respected 61 00:03:17,960 --> 00:03:20,880 Speaker 1: and loved, and you'll be the better offline legend of 62 00:03:20,880 --> 00:03:22,960 Speaker 1: the week if you can help me out here, maybe 63 00:03:22,960 --> 00:03:26,040 Speaker 1: even the legend of the year. I'm also looking at 64 00:03:26,280 --> 00:03:29,440 Speaker 1: any and all information related to data center costs. I 65 00:03:29,480 --> 00:03:32,160 Speaker 1: want to know the total cost of ownership of any AIGP, 66 00:03:32,320 --> 00:03:35,480 Speaker 1: and that includes the A one hundred, the H one hundred, 67 00:03:35,600 --> 00:03:37,680 Speaker 1: h two hundred, B one hundred, B two hundred, B 68 00:03:37,720 --> 00:03:41,960 Speaker 1: three hundred and so on. Blackwell obviously my biggest priority. 69 00:03:42,440 --> 00:03:45,320 Speaker 1: I want to know how many GPUs hyperscalers have too, 70 00:03:45,600 --> 00:03:48,000 Speaker 1: So if you work at Amazon, Microsoft or Google and 71 00:03:48,040 --> 00:03:50,560 Speaker 1: you know how many of the goddamn GPUs they've got, 72 00:03:50,680 --> 00:03:52,920 Speaker 1: please reach out. And I want to know the hourly 73 00:03:53,000 --> 00:03:56,400 Speaker 1: cost of running these GPUs, ideally on a per GPU basis. 74 00:03:56,560 --> 00:03:58,320 Speaker 1: I'm trying to work out how much it actually costs 75 00:03:58,320 --> 00:04:00,280 Speaker 1: to run these goddamn things, and it's in the same 76 00:04:00,360 --> 00:04:03,320 Speaker 1: we don't know now my Grail data. The things that 77 00:04:03,360 --> 00:04:06,400 Speaker 1: would materially change my reporting other than everything I just 78 00:04:06,480 --> 00:04:10,680 Speaker 1: mentioned would be the underlying cost of running large language models, 79 00:04:10,720 --> 00:04:13,680 Speaker 1: which means understanding both how many GPUs are used for 80 00:04:13,800 --> 00:04:17,440 Speaker 1: inference in training and the actual costs of running said GPUs. 81 00:04:17,960 --> 00:04:19,880 Speaker 1: The costs of running GPUs are at the center of 82 00:04:19,920 --> 00:04:21,599 Speaker 1: the bubble, and I believe that truth will be what 83 00:04:21,720 --> 00:04:25,560 Speaker 1: bursts in. Though I'm repeating myself a little. My other 84 00:04:25,600 --> 00:04:28,560 Speaker 1: grail data is the underlying cost of running CHET, GPT, Google, 85 00:04:28,640 --> 00:04:32,200 Speaker 1: Gemini or any other major popular LLLM service such as 86 00:04:32,279 --> 00:04:36,400 Speaker 1: Cursor or rep them the same gohoest for Microsoft, Meta, Amazon, Google, Oracle, 87 00:04:36,480 --> 00:04:40,359 Speaker 1: or really any other major cloud provider. Look together, I 88 00:04:40,360 --> 00:04:42,200 Speaker 1: believe we can make the world better by providing the 89 00:04:42,240 --> 00:04:45,120 Speaker 1: public with the truth. And this era is one steeped 90 00:04:45,120 --> 00:04:48,719 Speaker 1: in outright lines and stal selfishness that deprives good companies 91 00:04:48,760 --> 00:04:51,640 Speaker 1: of funding and shoves dysfunctional software on millions of people 92 00:04:51,839 --> 00:04:55,000 Speaker 1: that just wish it would go the fuck away. Now, 93 00:04:55,080 --> 00:04:57,280 Speaker 1: so many of you have already been so helpful, and 94 00:04:57,320 --> 00:04:59,479 Speaker 1: I hope I can I can hear from more of you. 95 00:05:00,080 --> 00:05:02,279 Speaker 1: The show continues to grow. It's been a crazy year. 96 00:05:02,560 --> 00:05:05,080 Speaker 1: I'm insanely grateful to have all of you. And next 97 00:05:05,080 --> 00:05:07,200 Speaker 1: week we have a three part in video podcast that 98 00:05:07,240 --> 00:05:10,360 Speaker 1: tells you all about the largest we'd just weirdest. I'm 99 00:05:10,400 --> 00:05:13,039 Speaker 1: not fixing it. I'm not fixing that you want the 100 00:05:13,120 --> 00:05:16,440 Speaker 1: raw stuff anyway. In video, it's the weirdest company in 101 00:05:16,480 --> 00:05:19,760 Speaker 1: stock market history. Oh God, I'm gonna hear from you 102 00:05:19,800 --> 00:05:22,400 Speaker 1: all about that one. I really do love hearing from 103 00:05:22,440 --> 00:05:26,320 Speaker 1: you though it's been really crazy. Twenty twenty five has 104 00:05:26,360 --> 00:05:29,240 Speaker 1: been crazy and bad and good in many different ways. 105 00:05:29,760 --> 00:05:32,000 Speaker 1: And I'm sure I'm gonna I'm gonna hear fun stories 106 00:05:32,000 --> 00:05:33,960 Speaker 1: from you all because a bunch of you reach out 107 00:05:33,960 --> 00:05:36,080 Speaker 1: and no matter what I say, to reach out about or 108 00:05:36,120 --> 00:05:38,040 Speaker 1: not reach out about, and I kind of love it. 109 00:05:38,200 --> 00:05:41,440 Speaker 1: I love how chatty you all are. So yeah, look 110 00:05:41,480 --> 00:05:43,840 Speaker 1: forward to hearing from you, and I hope you enjoy 111 00:05:43,920 --> 00:05:46,800 Speaker 1: the podcast about the Weirdest company on the stock market.