WEBVTT - Monologue: Silicon Valley is A Cargo Cult

0:00:02.440 --> 0:00:08.560
<v Speaker 1>Also media, Hello and welcome to this week's Better Offline Monologue.

0:00:08.600 --> 0:00:19.079
<v Speaker 1>I'm your host ed ze tron. It's gonna be a

0:00:19.120 --> 0:00:22.920
<v Speaker 1>double monologue week. After last week's dramatic exclusive, I've been

0:00:22.960 --> 0:00:25.720
<v Speaker 1>slammed with interviews and personal life stuff that made it

0:00:25.760 --> 0:00:27.160
<v Speaker 1>feel like I was being hit in the head with

0:00:27.200 --> 0:00:29.880
<v Speaker 1>a pool noodle twenty four to seven. Feeling a lot

0:00:29.960 --> 0:00:32.479
<v Speaker 1>better now, but nevertheless, this is where I need to

0:00:32.520 --> 0:00:35.600
<v Speaker 1>focus myself. I've also been enjoying a lot of the

0:00:35.640 --> 0:00:38.640
<v Speaker 1>responses to my exclusive about open Aiyes twenty twenty four

0:00:38.640 --> 0:00:42.600
<v Speaker 1>and twenty twenty five financials Now. Some have been mighty

0:00:42.680 --> 0:00:45.800
<v Speaker 1>confident about Inference being profitable due to a seven point

0:00:45.880 --> 0:00:49.120
<v Speaker 1>five billion dollar cost of revenue on thirteen point h

0:00:49.159 --> 0:00:53.159
<v Speaker 1>seven billion dollars in revenue, but overlooked my reporting from

0:00:53.240 --> 0:00:55.680
<v Speaker 1>last year verified by the Financial Times, by the way,

0:00:55.920 --> 0:00:58.560
<v Speaker 1>that showed open Aye spent eight point six seven billion

0:00:58.600 --> 0:01:01.040
<v Speaker 1>dollars specifically on infrared in the first nine months of

0:01:01.080 --> 0:01:04.640
<v Speaker 1>twenty twenty five. It's very very clear that open Ai

0:01:04.800 --> 0:01:07.200
<v Speaker 1>moved around numbers to make things look better than they are,

0:01:07.440 --> 0:01:09.560
<v Speaker 1>and I believe that inference costs are being dumped into

0:01:09.640 --> 0:01:12.679
<v Speaker 1>sales and marketing How else can you explain how a

0:01:12.720 --> 0:01:15.360
<v Speaker 1>company spends more than forty three percent of its revenue

0:01:15.440 --> 0:01:18.440
<v Speaker 1>five point seven three billion dollars on sales and marketing,

0:01:18.640 --> 0:01:21.640
<v Speaker 1>more than the Coca Cola Corporation, which has three ad

0:01:21.680 --> 0:01:24.240
<v Speaker 1>agencies and a vast web of different print, digital and

0:01:24.280 --> 0:01:26.640
<v Speaker 1>physical ads spend. I think they do some influence of

0:01:26.680 --> 0:01:29.920
<v Speaker 1>things as well. Hotously, if the Diet Coke Corporation wants

0:01:29.920 --> 0:01:33.840
<v Speaker 1>to get in touch, please do. Nevertheless, Microsoft also had

0:01:33.840 --> 0:01:36.399
<v Speaker 1>five hundred million dollars of sales and marketing spend too

0:01:36.480 --> 0:01:39.160
<v Speaker 1>via open Ai. What do you think that is open

0:01:39.200 --> 0:01:42.360
<v Speaker 1>ai spending five hundred million dollars on sales and marketing

0:01:43.000 --> 0:01:47.319
<v Speaker 1>through Microsoft? Or is it them itemizing promotional spend or

0:01:47.319 --> 0:01:51.000
<v Speaker 1>the inference and free users as a sales and marketing cost. Now,

0:01:51.040 --> 0:01:52.960
<v Speaker 1>if you disagree, and I'm sure some of you do,

0:01:53.240 --> 0:01:56.920
<v Speaker 1>please explain in any detail, love any level of detail, really,

0:01:57.200 --> 0:01:59.720
<v Speaker 1>how open ai has spent five point six seven billion

0:01:59.720 --> 0:02:03.600
<v Speaker 1>dollars and sales and marketing otherwise. Its first major advertising

0:02:03.600 --> 0:02:06.640
<v Speaker 1>campaign was in September twenty twenty five. If it's spending

0:02:06.880 --> 0:02:09.160
<v Speaker 1>let's say two hundred and fifty thousand dollars a year

0:02:09.200 --> 0:02:11.840
<v Speaker 1>on its five hundred sales stuff that's still only one

0:02:11.919 --> 0:02:14.959
<v Speaker 1>hundred and twenty five million dollars. Unless open ai is

0:02:15.000 --> 0:02:17.680
<v Speaker 1>one of the single largest accounts in digital advertising and

0:02:17.919 --> 0:02:19.799
<v Speaker 1>speaking with people all over the shop, I don't think

0:02:19.840 --> 0:02:22.320
<v Speaker 1>that's true. I think it's far more likely that there

0:02:22.320 --> 0:02:25.440
<v Speaker 1>are actual costs being hidden. And this is the kind

0:02:25.480 --> 0:02:28.560
<v Speaker 1>of thing a company does when it has utter loathing

0:02:28.800 --> 0:02:31.840
<v Speaker 1>for its investors, the media, and the general public, a

0:02:31.880 --> 0:02:34.919
<v Speaker 1>brazen attempt to bury costs to make things look better

0:02:34.919 --> 0:02:37.840
<v Speaker 1>for an audience that's directly incentivized to take any shred

0:02:37.919 --> 0:02:41.280
<v Speaker 1>of proof that things are okay. Nowhere is this more

0:02:41.320 --> 0:02:44.720
<v Speaker 1>obvious than in the FT's story, and I'm specifically referring

0:02:44.720 --> 0:02:47.120
<v Speaker 1>to a quote where a source familiar with the matter

0:02:47.200 --> 0:02:50.680
<v Speaker 1>suggested that somehow open ai only lost eight billion dollars

0:02:50.720 --> 0:02:53.200
<v Speaker 1>in twenty twenty five, rather than the twenty point nine

0:02:53.440 --> 0:02:57.280
<v Speaker 1>billion dollars it actually lost. Accountancy voodoo doesn't work on me.

0:02:57.400 --> 0:03:01.160
<v Speaker 1>Clam you well, I'm better than that. But this comment,

0:03:01.240 --> 0:03:03.440
<v Speaker 1>in my opinion, was not actually meant for the readers

0:03:03.480 --> 0:03:06.079
<v Speaker 1>of the FT, but for the larger AI cargo cult.

0:03:06.480 --> 0:03:09.160
<v Speaker 1>As I wrote in my free newsletter this week, Silicon

0:03:09.240 --> 0:03:12.520
<v Speaker 1>Valley as exemplified by the AI bubble, has abandoned its

0:03:12.560 --> 0:03:16.880
<v Speaker 1>supposed principles of rationality and meritocracy, and now more closely

0:03:16.880 --> 0:03:20.600
<v Speaker 1>resembles a cargo cult desperate to recreate bygone eras of growth.

0:03:21.000 --> 0:03:23.920
<v Speaker 1>Open AI and Anthropic are not interested in operating with

0:03:23.960 --> 0:03:26.880
<v Speaker 1>honesty or integrity, nor are they particularly concerned with building

0:03:26.919 --> 0:03:29.920
<v Speaker 1>real companies. They lose billions of dollars. They always have

0:03:30.000 --> 0:03:33.519
<v Speaker 1>and they always will. Nothing changes that. They must keep

0:03:33.800 --> 0:03:36.320
<v Speaker 1>the mythology of their company's going so that they can

0:03:36.400 --> 0:03:39.320
<v Speaker 1>keep raising money in perpetuity until someone works out how

0:03:39.320 --> 0:03:41.640
<v Speaker 1>to build a real company. Not that that will happen.

0:03:41.760 --> 0:03:45.200
<v Speaker 1>But nevertheless, the things that they say they exist to

0:03:45.240 --> 0:03:47.880
<v Speaker 1>feed the tech media in AI Booster's symbolic proof that

0:03:47.920 --> 0:03:50.960
<v Speaker 1>the great prophecy that Open Ai and Anthropic will grow

0:03:50.960 --> 0:03:53.400
<v Speaker 1>and be the biggest companies in the world. They'll be profitable,

0:03:53.400 --> 0:03:55.720
<v Speaker 1>and AI will be the next Big Three. Well, they

0:03:55.760 --> 0:03:58.920
<v Speaker 1>need that, They need you to believe these things so

0:03:58.960 --> 0:04:01.200
<v Speaker 1>that the great prophecy will come true. It must come true.

0:04:01.240 --> 0:04:04.360
<v Speaker 1>Don't you understand if it doesn't come true, Well, Warrio

0:04:04.480 --> 0:04:08.160
<v Speaker 1>Amiday can't be an elder statesman forever. And this is

0:04:08.160 --> 0:04:11.040
<v Speaker 1>why both companies have used annualized run rates and never

0:04:11.080 --> 0:04:15.680
<v Speaker 1>really addressed the profitability question outside of dario Amidae's stylized facts,

0:04:16.120 --> 0:04:19.320
<v Speaker 1>because they know that Silicon Valley is a monoculture dominated

0:04:19.360 --> 0:04:22.240
<v Speaker 1>by people who don't interact with society outside of sand

0:04:22.320 --> 0:04:25.599
<v Speaker 1>Hill Road, who they believe are gullible enough to accept

0:04:25.600 --> 0:04:29.080
<v Speaker 1>and promote any rationalization that confirms their beliefs or indeed

0:04:29.120 --> 0:04:33.120
<v Speaker 1>the great prophecy. And I mean they've they've been completely

0:04:33.200 --> 0:04:36.840
<v Speaker 1>right so far. The AI industry doesn't feel the need

0:04:36.880 --> 0:04:40.080
<v Speaker 1>to prove itself because it has complete contempt for its customers,

0:04:40.120 --> 0:04:43.039
<v Speaker 1>the media, and the markets, believing they'll slurp up whatever

0:04:43.080 --> 0:04:45.000
<v Speaker 1>slot they get and take it as proof that that

0:04:45.120 --> 0:04:48.440
<v Speaker 1>great prophecy, my friends, will come true and all of us,

0:04:48.640 --> 0:04:51.760
<v Speaker 1>all of us will be purifer And why wouldn't they.

0:04:52.400 --> 0:04:55.560
<v Speaker 1>The media is accepted basically every single narrative the AI

0:04:55.640 --> 0:04:59.599
<v Speaker 1>industry is created. AI boosters will accept effectively any proof

0:04:59.600 --> 0:05:02.599
<v Speaker 1>that things going well, and we'll actively attack and ostracize

0:05:02.600 --> 0:05:05.320
<v Speaker 1>any non believers. And if I'm honest, there are members

0:05:05.320 --> 0:05:07.080
<v Speaker 1>of the media who do this too. And if you're listening,

0:05:07.320 --> 0:05:10.680
<v Speaker 1>I'm not scared. You only make me stronger. You only

0:05:10.720 --> 0:05:14.839
<v Speaker 1>make my posts longer too. But really, this is how

0:05:14.839 --> 0:05:17.560
<v Speaker 1>it all works. This is how it's worked with AI.

0:05:17.600 --> 0:05:19.920
<v Speaker 1>It's how they tried to make crypto and the metaverse work.

0:05:19.960 --> 0:05:22.880
<v Speaker 1>But it really worked with AI because lllm's two do

0:05:22.960 --> 0:05:27.159
<v Speaker 1>that kind of cargo cultish evans generation. Oh, it could

0:05:27.200 --> 0:05:31.159
<v Speaker 1>potentially replace software engineers because it can burp out some code. Oh,

0:05:31.240 --> 0:05:33.640
<v Speaker 1>it can look at my taxes and say stuff that

0:05:33.680 --> 0:05:38.240
<v Speaker 1>means it can do my taxes, just enough symbology to

0:05:38.360 --> 0:05:42.159
<v Speaker 1>push this bullshit through for an indeterminate end. Let's be honest,

0:05:42.560 --> 0:05:44.279
<v Speaker 1>it's not going to be a government bailout. I've covered

0:05:44.360 --> 0:05:46.120
<v Speaker 1>that in a previous monologue. Go back and listen to it.

0:05:46.240 --> 0:05:51.240
<v Speaker 1>Not repeating myself, but this is Silicon Valley, and Silicon

0:05:51.320 --> 0:05:56.560
<v Speaker 1>Valley stands for rugged individualism, rationality, and meritocracy, unless, of course,

0:05:56.640 --> 0:05:59.120
<v Speaker 1>everybody in Silicon Valley agrees that idea is good, at

0:05:59.120 --> 0:06:01.800
<v Speaker 1>which point all that bullshit goes out the window. Any

0:06:01.839 --> 0:06:04.640
<v Speaker 1>singular piece of information that even slightly brings positive is

0:06:04.680 --> 0:06:07.320
<v Speaker 1>proof of the grand prophecy of growth that AI promises,

0:06:07.680 --> 0:06:10.799
<v Speaker 1>and chosen are those who take said proof, and cursed

0:06:10.920 --> 0:06:13.800
<v Speaker 1>are those who question it. You dare to report an

0:06:13.839 --> 0:06:18.479
<v Speaker 1>open aiy's actual revenues. No non believer help with him.

0:06:18.800 --> 0:06:20.880
<v Speaker 1>It's we're going to put him in the wicker Man.

0:06:20.920 --> 0:06:22.960
<v Speaker 1>We're going to put the man with the revenues in

0:06:23.000 --> 0:06:27.440
<v Speaker 1>the wicker Man. Except this is this is more like that. Yeah,

0:06:27.440 --> 0:06:29.320
<v Speaker 1>I guess this would be the wicked Man, except I'm

0:06:29.320 --> 0:06:33.479
<v Speaker 1>not like a sexless British policeman. They're more on that. Later,

0:06:33.520 --> 0:06:36.159
<v Speaker 1>I'll go deep down the wicked man hole. Nevertheless, this

0:06:36.240 --> 0:06:39.200
<v Speaker 1>is Silicon Valley. Now, Silicon Valley is just a cult.

0:06:39.279 --> 0:06:42.040
<v Speaker 1>It's a cargo cult where everyone's trying to shuffle shit around,

0:06:42.240 --> 0:06:44.599
<v Speaker 1>put things in the right order to try and usher

0:06:44.640 --> 0:06:47.000
<v Speaker 1>in the next Steve Jobs or the next Bill Gates.

0:06:47.200 --> 0:06:49.480
<v Speaker 1>You ever notice as well that that's all they talk about.

0:06:49.520 --> 0:06:52.599
<v Speaker 1>It's always the guys, it's never the product. You also,

0:06:52.640 --> 0:06:55.560
<v Speaker 1>ever notice that all of these extremely well paid AI

0:06:55.680 --> 0:06:59.520
<v Speaker 1>guys like a Nome shazeog or bought He left Google

0:06:59.520 --> 0:07:02.280
<v Speaker 1>and got bored for two point seven billion dollars by

0:07:02.480 --> 0:07:04.800
<v Speaker 1>Google to go back to Google, then went to open

0:07:04.839 --> 0:07:07.240
<v Speaker 1>ai a month ago. What do you think he did

0:07:07.240 --> 0:07:10.520
<v Speaker 1>at Google? What do you think Barrett Zaff, who left

0:07:10.560 --> 0:07:12.800
<v Speaker 1>open ai too, went to thinking machines that he came

0:07:12.840 --> 0:07:14.840
<v Speaker 1>back to open ai and knows left open Ai. What

0:07:14.880 --> 0:07:17.480
<v Speaker 1>do you think he's done? And now Andre Carpathy he

0:07:17.560 --> 0:07:20.120
<v Speaker 1>just went to Anthropic and he wasn't doing anything for

0:07:20.160 --> 0:07:23.000
<v Speaker 1>two years. Isn't it weird that Andre Carpathy, one of

0:07:23.000 --> 0:07:26.960
<v Speaker 1>the co founders of open Ai, formerly of Tesla, he

0:07:27.280 --> 0:07:29.520
<v Speaker 1>sat around for two years and really didn't do anything.

0:07:29.520 --> 0:07:32.760
<v Speaker 1>He had that Eureka Labs education startup, nothing happened with that.

0:07:33.600 --> 0:07:37.440
<v Speaker 1>It's almost as if the valley moves around different forms

0:07:37.480 --> 0:07:41.000
<v Speaker 1>of symbology hoping that when they're in the right place,

0:07:41.080 --> 0:07:44.280
<v Speaker 1>the next era of growth will begin. And doing this

0:07:44.400 --> 0:07:46.960
<v Speaker 1>allows them to get away from the horrible fact that

0:07:47.000 --> 0:07:50.160
<v Speaker 1>they're out of hypergrowth ideas, that they don't have the

0:07:50.200 --> 0:07:53.440
<v Speaker 1>people who have the ideas anymore. The original iPhone was

0:07:53.480 --> 0:07:56.880
<v Speaker 1>created by a competing teams Tony Fadell and John Forstall.

0:07:56.960 --> 0:07:59.480
<v Speaker 1>I believe the software and hardware teams coming together and

0:07:59.480 --> 0:08:03.760
<v Speaker 1>actually built something people with real industrial expertise. The valley

0:08:03.800 --> 0:08:07.720
<v Speaker 1>now is just I don't know, trading different kinds of

0:08:07.840 --> 0:08:10.640
<v Speaker 1>AI scientists and mathematician in the hopes that one of

0:08:10.680 --> 0:08:13.920
<v Speaker 1>them can usher in the digital god. It's fucking depressing.

0:08:14.880 --> 0:08:19.040
<v Speaker 1>It's depressing, and make no mistake, it will lead Silicon

0:08:19.120 --> 0:08:22.080
<v Speaker 1>Valley to its ruin. Faith can't fill your belly, or

0:08:22.120 --> 0:08:24.720
<v Speaker 1>cover one point one trillion dollars in spending commitments, or

0:08:24.800 --> 0:08:27.960
<v Speaker 1>magic up the more than two trillion dollars in revenue

0:08:28.000 --> 0:08:31.000
<v Speaker 1>that big tech needs to make AI worth it both

0:08:31.160 --> 0:08:33.800
<v Speaker 1>twenty thirty. I don't think it can paid Larry Ellison's

0:08:33.800 --> 0:08:37.240
<v Speaker 1>bills either. In fact, faith can't really do much other

0:08:37.280 --> 0:08:41.480
<v Speaker 1>than create speculative wealth, and speculative wealth can't pay for

0:08:41.559 --> 0:08:45.240
<v Speaker 1>everything forever. I'll got you on Friday. This has been

0:08:45.280 --> 0:08:47.400
<v Speaker 1>better offline, and I bet ed Zetron