WEBVTT - Monologue: ChatGPT's Growth Is Collapsing In Europe

0:00:02.440 --> 0:00:08.160
<v Speaker 1>Alzo Media. Hello and welcome to this week's Better Offline Monologue.

0:00:08.160 --> 0:00:18.520
<v Speaker 1>I'm your host ed ZiT Tron and sorry that there

0:00:18.520 --> 0:00:21.120
<v Speaker 1>hasn't been a monologue for a few weeks. These massive

0:00:21.120 --> 0:00:23.119
<v Speaker 1>episodes have been putting together, took up my time in

0:00:23.160 --> 0:00:25.520
<v Speaker 1>a way that made them a little impractical, and as

0:00:25.520 --> 0:00:27.520
<v Speaker 1>far as the show goes, I know it's been very

0:00:27.560 --> 0:00:31.240
<v Speaker 1>AI focused. I can't even say AI right sometimes. But anyway,

0:00:31.920 --> 0:00:34.760
<v Speaker 1>this is likely going to continue for a while because,

0:00:34.840 --> 0:00:37.080
<v Speaker 1>as you can tell from the news, crazy shit keeps

0:00:37.080 --> 0:00:40.319
<v Speaker 1>happening every day. Every day Sam Altman tells somebody he's

0:00:40.320 --> 0:00:42.920
<v Speaker 1>going to give them a billion dollars. It's all so silly,

0:00:43.280 --> 0:00:45.640
<v Speaker 1>and I do feel like it's not being covered very thoroughly,

0:00:45.760 --> 0:00:48.960
<v Speaker 1>so I'm doing my best to do so. Nevertheless, I

0:00:49.200 --> 0:00:51.000
<v Speaker 1>promise I'm going to find a way to bring in

0:00:51.040 --> 0:00:54.680
<v Speaker 1>more non AI stuff. I'm not really sure how it's

0:00:54.720 --> 0:00:56.640
<v Speaker 1>going to make it in there. I'll do my best,

0:00:56.680 --> 0:00:59.360
<v Speaker 1>and I'm already working on CES twenty twenty six, which

0:00:59.400 --> 0:01:01.319
<v Speaker 1>I think is going to be probably the best tech

0:01:01.360 --> 0:01:05.840
<v Speaker 1>podcasting thing ever. Got some incredible guests Chloe Radcliffe of course,

0:01:05.959 --> 0:01:09.120
<v Speaker 1>Edward Anguiso Junior, David Roth, and yeah, those guests I

0:01:09.160 --> 0:01:10.880
<v Speaker 1>can't quite talk about yet, but they're gonna make you

0:01:10.920 --> 0:01:12.839
<v Speaker 1>really happy, and it's gonna be about more than just AI.

0:01:13.640 --> 0:01:16.600
<v Speaker 1>That being said, next week's episode is likely to be

0:01:16.640 --> 0:01:19.240
<v Speaker 1>AI focused, as I've had a lot of people request

0:01:19.240 --> 0:01:21.360
<v Speaker 1>they do an episode that's a simplified guide to the

0:01:21.400 --> 0:01:24.039
<v Speaker 1>AI bubble, something that they can share with people that

0:01:24.040 --> 0:01:26.600
<v Speaker 1>aren't sticking their head and the technological or financial septic

0:01:26.640 --> 0:01:30.000
<v Speaker 1>tank every day. And look, I think you're gonna love it.

0:01:30.520 --> 0:01:33.039
<v Speaker 1>You can please listen to it anyway. I need the downloads,

0:01:33.080 --> 0:01:35.720
<v Speaker 1>but in all seriousness, I think it's gonna be something,

0:01:35.760 --> 0:01:37.600
<v Speaker 1>and I've been working on it for a few days now.

0:01:37.800 --> 0:01:40.880
<v Speaker 1>That's quite short and sharp, but we'll speak to people

0:01:40.920 --> 0:01:43.600
<v Speaker 1>who are not technical or financial at all. I think

0:01:43.600 --> 0:01:46.800
<v Speaker 1>it'll be great and if not shows free you'll enjoy it.

0:01:46.840 --> 0:01:49.160
<v Speaker 1>But I also do just want to say I really

0:01:49.200 --> 0:01:51.960
<v Speaker 1>appreciate how many of you listen every week. I really

0:01:52.000 --> 0:01:55.040
<v Speaker 1>appreciate your patience with the AI stuff. You seem to

0:01:55.080 --> 0:01:57.520
<v Speaker 1>all really enjoy it. But I do love to hear

0:01:57.520 --> 0:02:00.400
<v Speaker 1>from you whenever I can, so reach out now. Please

0:02:00.440 --> 0:02:02.120
<v Speaker 1>don't be too mean, though I do have a heart,

0:02:02.520 --> 0:02:06.120
<v Speaker 1>But yes, this Week's monologue is somewhat of a big story.

0:02:06.920 --> 0:02:09.480
<v Speaker 1>So two days ago Fortune published a story saying that,

0:02:09.520 --> 0:02:11.680
<v Speaker 1>according to dates from Deutsche Bank, which is a very

0:02:11.760 --> 0:02:15.160
<v Speaker 1>very large bank, that OpenAI subscription growth in Europe has

0:02:15.160 --> 0:02:17.960
<v Speaker 1>stalled and is and I quote flatlining in the major

0:02:18.000 --> 0:02:21.799
<v Speaker 1>European markets over the past four months. To quote Fortune,

0:02:21.880 --> 0:02:24.799
<v Speaker 1>the data was drawn from transactions processed by third party

0:02:24.840 --> 0:02:29.520
<v Speaker 1>financial institutions. Really nailed that word in the UK, Germany, France,

0:02:29.600 --> 0:02:33.040
<v Speaker 1>Italy and Spain, representing fifteen percent of Chat GPT users.

0:02:33.440 --> 0:02:35.760
<v Speaker 1>And if you're wondering, the top software markets in the

0:02:35.800 --> 0:02:39.240
<v Speaker 1>world are US, China, Germany, UK, France, and Italy, and

0:02:39.280 --> 0:02:41.640
<v Speaker 1>then there's Canada and some others. To be clear, this

0:02:41.680 --> 0:02:45.040
<v Speaker 1>specifically refers to people paying for Chat GPT subscriptions the

0:02:45.080 --> 0:02:47.360
<v Speaker 1>twenty buck a month one most likely, and in a

0:02:47.400 --> 0:02:49.320
<v Speaker 1>graph I'll link to in the show notes, you really

0:02:49.360 --> 0:02:52.160
<v Speaker 1>should look at it. Open Ai subscription growth in Europe

0:02:52.200 --> 0:02:54.520
<v Speaker 1>has been pretty rocky from the beginning, dropping from ten

0:02:54.520 --> 0:02:57.400
<v Speaker 1>percent in June twenty twenty three to below zero percent

0:02:57.440 --> 0:03:00.359
<v Speaker 1>in August twenty twenty three, meaning growth to decline only

0:03:00.400 --> 0:03:02.840
<v Speaker 1>to recover slightly well pretty well, I guess the twenty

0:03:02.840 --> 0:03:05.160
<v Speaker 1>percent in December twenty twenty three, then drop again to

0:03:05.200 --> 0:03:08.920
<v Speaker 1>blow zero percent in February twenty twenty four. Otherwise, open

0:03:08.960 --> 0:03:11.920
<v Speaker 1>Ai barely scraped ten percent growth outside of some spikes

0:03:11.919 --> 0:03:14.400
<v Speaker 1>in September twenty twenty four around the launch of their

0:03:14.400 --> 0:03:16.960
<v Speaker 1>Reasoning models which got them to about fifteen percent, the

0:03:17.040 --> 0:03:20.200
<v Speaker 1>launch of their Useless Agent operator in January twenty twenty five,

0:03:20.240 --> 0:03:23.480
<v Speaker 1>which got them somehow to around eighty eighteen percent, and

0:03:23.480 --> 0:03:26.320
<v Speaker 1>then the launch of four Image Generation in March twenty

0:03:26.320 --> 0:03:29.040
<v Speaker 1>twenty five, which by the way was the studio Ghibliwank,

0:03:29.280 --> 0:03:31.720
<v Speaker 1>and that got them back to twenty ten percent, though

0:03:32.240 --> 0:03:34.520
<v Speaker 1>I realized that was a lot of numbers, and like

0:03:34.560 --> 0:03:37.240
<v Speaker 1>I said, the chart is in the episode Notes I swear.

0:03:37.440 --> 0:03:40.400
<v Speaker 1>But in June twenty twenty five, open Ai faced negative

0:03:40.440 --> 0:03:43.360
<v Speaker 1>growth of subscribers in Europe, barely recovering to two to

0:03:43.400 --> 0:03:45.880
<v Speaker 1>three percent in July twenty twenty five, and then back

0:03:45.880 --> 0:03:48.920
<v Speaker 1>to zero percent in August twenty twenty five, which looks

0:03:48.960 --> 0:03:51.520
<v Speaker 1>like they're maybe a mild recovery, coming to around three

0:03:51.560 --> 0:03:56.080
<v Speaker 1>percent in September. Now, while I'm not quite declaring victory

0:03:56.200 --> 0:03:58.880
<v Speaker 1>or blasting not like us, I am having trouble finding

0:03:58.880 --> 0:04:01.840
<v Speaker 1>a way around how bad. I can find no example

0:04:01.840 --> 0:04:04.520
<v Speaker 1>of a successful software revolution that started with high growth

0:04:04.560 --> 0:04:07.080
<v Speaker 1>and tumbled into the abyss in Europe. It didn't happen

0:04:07.120 --> 0:04:10.320
<v Speaker 1>with Facebook, and it certainly didn't happen in Uber's case,

0:04:10.440 --> 0:04:13.400
<v Speaker 1>even though European governments actively tried to reject it, and

0:04:13.480 --> 0:04:16.240
<v Speaker 1>some of them still are. It didn't happen with the smartphone,

0:04:16.279 --> 0:04:19.320
<v Speaker 1>or cloud computing, or really any other software revolution or device.

0:04:19.640 --> 0:04:22.239
<v Speaker 1>They would start from kind of nowhere and then grow,

0:04:22.440 --> 0:04:25.000
<v Speaker 1>and then there'd be wobbles, but they'd keep growing steadily.

0:04:25.320 --> 0:04:28.359
<v Speaker 1>This is a steady decline. And to be clear, this

0:04:28.400 --> 0:04:31.600
<v Speaker 1>has nothing to do with Europe's General Data Protection regulations

0:04:31.600 --> 0:04:34.000
<v Speaker 1>of the GDPR. Robert Evans once told me to spell

0:04:34.040 --> 0:04:37.159
<v Speaker 1>everything out. I've independently confirmed with the European chet GPT

0:04:37.400 --> 0:04:40.360
<v Speaker 1>users that the product is basically identical over there. Europe

0:04:40.400 --> 0:04:43.880
<v Speaker 1>just happens to be rejecting chet GPT. They don't want

0:04:43.880 --> 0:04:45.520
<v Speaker 1>to pay for it. And as you'll see when you

0:04:45.520 --> 0:04:47.880
<v Speaker 1>look at the graph, the boosts that open Ai gets

0:04:47.920 --> 0:04:51.559
<v Speaker 1>from throwing out new products, they're becoming less pronounced each time.

0:04:51.880 --> 0:04:54.440
<v Speaker 1>Other than operator that one kind of surprised me but

0:04:54.480 --> 0:04:58.120
<v Speaker 1>even then eighteen percent growth, and I think it's month

0:04:58.200 --> 0:05:00.560
<v Speaker 1>over month. If it's year over year, they wouldn't be

0:05:00.560 --> 0:05:03.000
<v Speaker 1>a year of year, it's month over month. Now. The GDPR,

0:05:03.040 --> 0:05:04.800
<v Speaker 1>by the way, will make it difficult for open ai

0:05:04.839 --> 0:05:07.359
<v Speaker 1>to launch advertising, which I will add will not be

0:05:07.440 --> 0:05:09.960
<v Speaker 1>the saving grace for this company, because Google and Meta's

0:05:10.040 --> 0:05:13.600
<v Speaker 1>massive advertising revenues come from decades of data and expertise

0:05:13.640 --> 0:05:16.280
<v Speaker 1>in forcing them down customers throats, as well as having

0:05:16.279 --> 0:05:19.760
<v Speaker 1>a very compelling product and thousands of salespeople and also,

0:05:19.800 --> 0:05:22.200
<v Speaker 1>in Google's case, thousands of miles of underground cable and

0:05:22.279 --> 0:05:24.560
<v Speaker 1>under the c cable. I believe as well people that

0:05:24.560 --> 0:05:27.039
<v Speaker 1>are easily scared of suggesting the open AI's data on

0:05:27.160 --> 0:05:29.240
<v Speaker 1>users is rich because they feed so much into it.

0:05:29.520 --> 0:05:32.080
<v Speaker 1>This point makes the assumption that this data is useful

0:05:32.320 --> 0:05:34.560
<v Speaker 1>and that open ai can actually do anything with it,

0:05:34.600 --> 0:05:38.000
<v Speaker 1>despite having absolutely no expertise in selling ads. As a

0:05:38.000 --> 0:05:41.200
<v Speaker 1>reminder perplexity that an AI search engine, you'd think would

0:05:41.200 --> 0:05:43.760
<v Speaker 1>be the perfect place for ads, they made twenty grand,

0:05:44.360 --> 0:05:48.799
<v Speaker 1>twenty thousand dollars selling ads in twenty twenty four. That's bollocks.

0:05:48.839 --> 0:05:52.200
<v Speaker 1>It's terrible. You make more selling dogs. They will watch

0:05:52.240 --> 0:05:54.760
<v Speaker 1>the day to day might be auctioning dogs. Anyway, my

0:05:54.880 --> 0:05:57.560
<v Speaker 1>rambling aside, it's very easy to get scared of this

0:05:57.600 --> 0:05:59.360
<v Speaker 1>company and just think, oh, they have a lot of

0:05:59.400 --> 0:06:01.960
<v Speaker 1>money in you. Even though they lose billions, they'll just

0:06:02.000 --> 0:06:04.880
<v Speaker 1>put ads in and that'll fix everything. I must caution

0:06:05.080 --> 0:06:08.560
<v Speaker 1>you to immediately be a doomer there. I see no

0:06:08.720 --> 0:06:11.640
<v Speaker 1>real evidence that they're going to find another business line

0:06:12.000 --> 0:06:14.240
<v Speaker 1>in mostly evidence by the fact they've never been able

0:06:14.279 --> 0:06:17.039
<v Speaker 1>to find one. And while this isn't declarative proof that

0:06:17.080 --> 0:06:19.600
<v Speaker 1>this company is circling the drain, I don't know how

0:06:19.640 --> 0:06:21.880
<v Speaker 1>else to look at it as anything other than their

0:06:21.920 --> 0:06:24.760
<v Speaker 1>growth story in Europe is dying and they're building a

0:06:24.760 --> 0:06:27.440
<v Speaker 1>goddamn data center in England. I think with n Scale,

0:06:27.440 --> 0:06:29.360
<v Speaker 1>the company that's never built an AI data center in

0:06:29.400 --> 0:06:31.960
<v Speaker 1>their life. To be fair, that's also the case with Crusoe.

0:06:31.960 --> 0:06:35.440
<v Speaker 1>In fact, all of Openay's partners pretty much never built

0:06:35.480 --> 0:06:38.000
<v Speaker 1>a data center of any kind other than crypto ones,

0:06:38.000 --> 0:06:40.560
<v Speaker 1>which are completely different. Also, no one's built a gig

0:06:40.600 --> 0:06:43.080
<v Speaker 1>what data center. Ah, that's going to be a future episode.

0:06:43.120 --> 0:06:45.400
<v Speaker 1>Forgive me. I want to educate you on data centers

0:06:45.440 --> 0:06:47.480
<v Speaker 1>one day, but every time I start thinking about them.

0:06:47.520 --> 0:06:51.280
<v Speaker 1>My brain starts playing that music from kill Bill when

0:06:51.360 --> 0:06:53.760
<v Speaker 1>she looks anyone she's mad at. Someone's going to email

0:06:53.800 --> 0:06:56.920
<v Speaker 1>me and say, Perry Mason Christ. The great thing about

0:06:56.920 --> 0:06:59.640
<v Speaker 1>these monologues is I could just talk about whatever I

0:06:59.640 --> 0:07:01.560
<v Speaker 1>want to. They say I should do this, which is

0:07:01.600 --> 0:07:05.080
<v Speaker 1>what I'm gonna do. But in reality, back to the

0:07:05.080 --> 0:07:07.880
<v Speaker 1>thing we're talking about, it really is hard to read

0:07:07.920 --> 0:07:10.480
<v Speaker 1>this as anything other than air escaping the bubble. It's

0:07:10.520 --> 0:07:13.440
<v Speaker 1>been pissed, I swear to God, and perhaps it's the

0:07:13.600 --> 0:07:17.040
<v Speaker 1>error of smiles beginning. A specter is haunting Silicon Valley,

0:07:17.280 --> 0:07:20.280
<v Speaker 1>the face of the smiling man, and I hope you'll

0:07:20.360 --> 0:07:22.160
<v Speaker 1>join me every week to find out how this shit

0:07:22.280 --> 0:07:25.960
<v Speaker 1>bursts or I don't know, whatever crazy crap Sam Mortman

0:07:26.000 --> 0:07:26.800
<v Speaker 1>promises next