The Vergecast - What's behind the Google AI shakeup

Starting point is 00:00:00 Support for the show comes from Atio, the CRM for teams who set the pace. Imagine if your CRM could learn your business and build itself around it. That's exactly what you get with Atio, the agentic CRM. All you have to do is connect your inbox and calendar, then every lead gets scored for fit. You get full briefs in hand before recall, and personalized follow-ups in your voice, get drafted to target accounts. You can go to atio.com slash vergecast, and you'll get 15% of it. off your first year. That's A-T-T-I-O.com slash Vergecast. Support for the show comes from ServiceNow.

Starting point is 00:00:41 AI is moving fast across the enterprise, but without visibility, it's just chaos. Different tools, different models, different teams using AI in completely different ways. Service Now turns that chaos into control. With the AI control tower, you see all your AI across the business in one place. what it's doing, what it's done, and what it's about to do, so you stay in control. To put AI to work for people, visit servicenow.com. Thank you for calling the Bombas Comfort Line. Bombas make socks, slippers, teas, and underwear made with the highest quality materials.

Starting point is 00:01:19 Press 1 for comfort. 2 for style. 3 for donation. You chose style. Bombas is styles for whatever you enjoy. You can run in bambas, lounge in bambas, dress them up, dress them down, but always give back in bambas. because with every item purchased, another is donated. Bambas, comfort worth calling for.

Starting point is 00:01:36 Go to bambas.com slash audio and use code audio for 20% off your first purchase. That's BOMBAS.com and use code audio. Hello and welcome to the Vergecast, the flagship podcast of ticking fees. For every minute, this podcast goes, Nilei owes me $1 million. I'm your friend David Pierce, Neil Iber tells here. Hello. I won't pay. That works in the current state of things.

Starting point is 00:01:59 It's actually a pretty good strategy. Did you read that there's an airtime story today about a town in California that had a recall election for the mayor and three county commissioners and they just won't leave? They're just like, yeah, we don't believe. Something illegal must have happened and they just didn't leave office. It's very good. That rocks, man. You can just do whatever you want and you'll probably get away with it. Right now, it's like, yeah, it's fine, whatever you want.

Starting point is 00:02:23 I own TikTok now. Let's go. Who does own TikTok now? Oh, we're going to get to it. We're coming to it. Yeah. It's going to be a part of this episode. Yeah. All right. We actually, we have a lot to talk about. We have a bunch of earnings stuff with some streaming news. The everybody wants to be YouTube phenomenon rages on. We have a lot of Paramount news. We have some car news.

Starting point is 00:02:44 Nelai is going to have a lot of feelings about BMW at some point in the show, I think, which I'm very much looking forward to. But we should start with the news from Wednesday about Google's AI leadership shakeup, right? This feels to me like kind of the biggest newsy story of the week. Oh, I think it is the story of the week. Okay. I want you to explain why, but let me just quick. explain. So two things happened that my understanding is are happening simultaneously, but not because of one another, if that makes sense. So on the one hand, you have Jeff Dean, who was, I think, employee number 30 at Google, along with a bunch of other respected AI researchers at Google, going off and starting their own company. They are doing what a lot of people in AI seem to be doing right now, which is realizing, oh, my God, there's so much money in this space. Why am I doing

Starting point is 00:03:29 this at a large tech company. We're going to go start a startup. Presumably they will raise tons of money and that will be that. Right next to it, you have a big shakeup inside of Google DeepMind, which is the team that runs all of the AI stuff inside of Google. If you remember, one of the reasons Google DeepMind is the way that it is is because Jeff Dean lost a fight to be in charge of AI at Google. In a very straightforward way, Demis Asab is sort of elevated to the top of all of Google's AI stuff. Jeff Dean did not. I think it is not entirely surprising that Jeff Dean after this time no longer works at Google. But anyway, Demis DeSavis, who has been running Google DeepMind for some time,

Starting point is 00:04:05 is moving to a different job. He's now the chief scientist at Alphabet and has, I think, is the chair of Google DeepMind now, some sort of honorific title that is basically he does not want to run the thing anymore, but he wants to keep going and being fancy and doing AI things. One of his sort of deputies, this guy named Kari, is now running Google DeepMind. And there's a sense going around with all of this news that something is a miss or maybe just confused or wrong or in some ways very Googly with Google's whole AI. So tell me what you make of all of this news. First of all, you know I love an org chart story, David.

Starting point is 00:04:47 I do. I got a whole podcast over there that's basically like, tell me about your org chart and I'll tell you everything that's wrong with you. And one of the cliches about org charts and products is you can look at a product and, you can look at a product and, you always ship your orchard. Yes. So if you look at Microsoft Word, you can see the political fights inside of Microsoft that led to all those buttons being there over time.

Starting point is 00:05:06 And I would just say you can look at Google's suite of AI products and realize that there's no cohesion in what Google is doing with AI. And I'll just give you three examples. I have a GM car that has Google built in it, and it just got upgraded to Gem. which sometimes goes away. This is probably GM's fault. But it's somewhere between Google and GM.

Starting point is 00:05:34 Sometimes it just goes away and becomes Google Assistant again. Weird. That's just a weird thing that happens in my car. Sometimes the all-powerful AI just disappears. They get tired. What are you going to do? Gemini in my car is supposed to be the same as Gemini in my phone. Right.

Starting point is 00:05:49 But it actually has a slightly different set of capabilities. And none of that is related to Google Assistant slash Gemini on my Google Home devices. in my home. Right. And all of that is different than whatever appears in Gmail and Google workspace. So today I asked my Google home,

Starting point is 00:06:08 hey, I got an email this morning. Can you find the, like, order code for me? And it was just like, I can't read your email. And I was like, the whole point of this is that you're supposed to be able to read my email. Like, no, you super can, Gemini. Why did I turn voice match on and all this stuff and give you access to my email

Starting point is 00:06:23 if I can't just say, hey, can you pull the order number out of my email for me? Right. And it could not do that. the best part is randomly, like, your car might have been able to. Do you know what I mean? Like, this is the Gemini story. Right. And this is all I'm saying is I can look at that mess of products. And they've put AI everywhere across the Google suite at really high rates. And they're very proud of this.

Starting point is 00:06:43 Yeah. And in many ways, they're winning, right? They're shipping more AI products to more people than anyone because they just have the most surface area. And you can say, well, there's no strategy here. There's no vision for how all of this makes a Google ecosystem powered by AI. there's just different riffs on Gemini in every single place they can put them. And, you know, Sundar was on Decoder at I.O. And his whole point of view was we've got the big foundational technology team that's Google DeepMind. And then we're productizing it all these different ways. And I get the sense that the product that Demis Asabas wanted to make a Google DeepMind was AGI. This is Demis saying over and over again, we're in the foothills of the singularity.

Starting point is 00:07:23 A phrase Sundar has now adopted, which makes it a thing, which sucks. It was in the blog post that, like, he's just parroting this as like, this is now, this is now the phrase we're using. It's like, live, laugh, love, the foothills of the singularity. Like, this is what we're doing now. Well, if you think we're in the foothills of the singularity, by the way, which generally defined is a point by which you cannot see through or beyond. Who knows, man, that AGI is going to come about and the super robots are going to be smart enough to do everything in the world and we'll just live a life of carefree, fully automated communism, whatever. If you think that's happening, you should spend all of your time focused on that. You should spend zero percent of your time allocating GPUs to Gmail search, which was part of Demis's job, right?

Starting point is 00:08:05 Like just fighting over who gets what access to what model when. And I think he wants to do automated drug discovery and bleeding edge frontier research. There's no question that Google is no longer at the frontier competing with open and anthropic, even on the latest benchmarks. They're way behind. Demis wants to go do that. And the very prosaic, let's put the Google assistant here powered by Gemini over there connecting to your data source over here to make real products people might like and not hate AI. That seems like someone else's problem. And I don't know, I don't know the new person who's going to run DeepMind.

Starting point is 00:08:41 It's notable that they're not a CEO. They're an SVP in Google structure. DeepMind is now a business unit inside of Google that has an SVP, a chip. and it doesn't have a CEO. It doesn't have business results to contend with on its own. So it's getting subsumed more into Google. And all of the researchers inside of Google, Hayden has a great story about this, are pretty mad that Google is doing military applications with AI.

Starting point is 00:09:09 So there's just a lot of turmoil inside of this organization that I think is reflected in the products. By the way, the products are the ones that have to make money while Demis goes and tries to escape the foothills to climb the mountain of the singularity. What do you think? I agree with almost all of that. I think the other piece of this that I've been thinking a lot about is where AI is interesting and where AI is actually finding real product market fit. Right?

Starting point is 00:09:38 Like if you look at Demis's office, I've talked to him a number of times over the years. And one of my favorite things about talking to Demis is you talk to Google executives or executives at almost any company. and you usually talk to them in the course of talking about some new product, right? They're about to launch a thing, so they make Demisis Abbas available to talk about that thing. He never wants to talk about the news. He's like, look at this cool new thing we shipped. I actually don't care very much about that. Let me tell you all these other things that we're working on and all the reasons I think this is interesting.

Starting point is 00:10:07 And he has been pretty straightforward over the years about not thinking that building an AI personal assistant to search your Gmail is the most interesting thing you can do with AI. I happen to think he's right. And he really believes, I mean, he won a Nobel Prize for some of the work that DeepMind has been doing for many years on this stuff to actually go and try to cure diseases and come up with new proteins and invent whole new things. Like, he just earnestly believes that all of the places we are pointing AI right now are the, are the uninteresting places to put it. And especially you look at it right now, we're in earnings season, so we're learning how much

Starting point is 00:10:48 all of these companies are spending on this stuff. Everybody's capital expenses are out of control. Everybody is spending like the world is going to end on AI infrastructure of one sort or another. And the only way any of that stuff is going to return for itself is business services, right? Like this is what we always talk about. AI is business software. The business of AI is business software. And again, if you're Dennis and you're really interested in curing cancer, like building B2B

Starting point is 00:11:14 AI infrastructure to make it slightly easier for people to do accounting at their companies is not going to be very interesting to you. And there's been some really good reporting since this news in semaphore and elsewhere that Demis's Havas has actually been kind of disengaging over the course of the last year because he doesn't want to run a giant organization. He doesn't enjoy the politics of Google by all accounts. He doesn't care about any of this. He's like, he is a true believer in, I think, the truest sense of what AGI is going to be and what it could be and what it could mean. And I think it's exactly what you said. if you are all the way at the end of that spectrum,

Starting point is 00:11:51 the hell are you doing working on business software? And I think the thing about this that I really enjoy is, my joke about AI all the time is AI is going to cure cancer way before it builds you a good personal assistant. You know who believes that is Demis Hasopus. And again, if that's what you believe, of course you should give yourself a giant fancy title like chief scientist at Alphabet that lets you go be on stages around the world to proclaim what AI is going to be.

Starting point is 00:12:20 It lets you go be influential all over Google instead of just a deep mind. It lets you not spend all of your time dealing with people fighting for compute at Google Maps versus Gmail. And it lets you go invent AGI. Like, is that a weird decision that's going to make him do a lot of odd things with his time over the next few years? Like, yes, absolutely. But it makes perfect sense to me. And I think is like pretty clearly where this is headed. So they're going to replace him with somebody who is a much more normal operator of a business unit inside of Google,

Starting point is 00:12:54 who will spend all of his time talking about the new features of Gemini, and that will be the job. I don't disagree with any of that. I think that is going to be fundamentally dangerous to a company like Google that does not have any top-down product strategy. Yes. And who is losing the person in Demisis Sassavis, who is maybe the single most credible. person in the entire AI industry. Yeah, we saw in the Musk v. Altman trial, they're all afraid of Dennis. Yeah, he is the boogeyman to every one of their competitors.

Starting point is 00:13:24 And he had the most resources and was in control of the most things. And I think part of having the most resources and being controlled the most things is you're also accountable to most decisions. And at Google, that means you're allocating GPUs. Right. Even while, you know, Anthropic is showing up and taking up Google Cloud capacity because Anthropic means all the capacity in the world. I think one of the most interesting things here is Google is not computer.

Starting point is 00:13:46 at the frontier with Anthropic and Open AI. I don't think that's true, by the way, but we can come back to that. I will come back to that. I mean, I just look at the benchmark charts. It's however much you want to believe the benchmarks. But Corai Kovachulu, I think is how you pronounce his last name. He's the former CTO of DeepMind. He will now become the SVP of DeepMind who will report to Sundar Pichai.

Starting point is 00:14:05 And the idea that he will also run the research lab, which is what DeepMind is supposed to be and make it a business unit that feeds into everything else. That's what Google has to unwind. And this is what Sooner was describing to me, you know, on that episode of the decoder at I.O., which is DeepMind is the big research facility, creates all the core capabilities, and then you go out and get productized.

Starting point is 00:14:28 And I think of your Dennis Isabas, you're like, no, the product is AGI. Yes, 100%. And if your job is to make the frontier models and those get front-productized elsewhere, it just puts you in a different position than Anthropic, which is, you know, every single day, their product is the model.

Starting point is 00:14:45 in like big meaningful ways. And they're not really making any money and they sometimes adjust their earnings to make it sound like they're making money. Open AI isn't making a dime. Their product is the model too. And I think Google just has to make a big decision about what the product is.

Starting point is 00:15:00 And right now they make money, even though they're spending tons and tons of money out in the world. If they stopped doing that, they would just start printing money again. Yeah. Like this last quarter, I think they posted a net cash operating loss because they're spending so much money in data.

Starting point is 00:15:15 Right. They could just stop. They could just stop doing it. And the only reason you're doing that is because you believe that there will be infinite demand for AI, for your AI, or even a solid capacity to entrop and open AI. And then Dennis is going to get to AGI. He's going to chase this all the way down. And then that will unlock whatever Larry Page and Sergey Brin will ascend to the next, even the next level of sentience that they have long dreamt of. And I, all that's a lot of bets, man. It is.

Starting point is 00:15:43 And I think if the products aren't good along the way, all of that starts to teeter. Okay, but this comes back to is Google behind? Right? Because one way to look at all of this news is Google flailing because it's losing, right? Gemini 3.5 was supposed to come out in June. Now it's August. It's not out yet. There has been some reporting that Gemini 3.5 is not up to the standards that Google was hoping for.

Starting point is 00:16:07 Jeff Dean is leaving. Demis Sassabas is leaving, like, full panic everywhere. is one way to look at all of this news. Another way to look at all of this news is actually Google continues to be in the strongest position in the AI industry, and it's not super close. And maybe it doesn't matter if Google has the best model because it's going to win in every single other imaginable way. It has the distribution. It has the vertical integration that it needs. It has the sales team that it requires.

Starting point is 00:16:36 Like, Google can probably win without having the best model. but I do think among a set of people, particularly in Silicon Valley, there is a real sense that Google is behind. And how you look at what's happening right now really depends on whether you think Google is behind or not. And it sounds like you do think Google is behind. I, again, the bleeding edge of the frontier, I can't evaluate that. I think a lot of people have a lot of opinions about that. And many of those people seem to think Google is behind. In terms of the things people are doing every day, the experiences people are,

Starting point is 00:17:09 having with AI that makes them hate AI and protest data centers. I don't think Google has as good of a product in AI overviews as free chat Shoe-E-T or free Claude. And we can see it every day because AI overviews are extraordinarily public. People keep using them and posting screenshots of how stupid they are. We have one here. There's a really good one from this week. So I'll just offer you this one, which is I think my funniest bad AI overview in quite some time. Okay. Bill Gruskin on Blue guy posted, he would clearly search for what is buying the dip means. And AI overrease spat out to him, buying the dip means purchasing an investment like Dean's French onion dip or stocks after a price dip. And then someone went and checked it. It's real. And then

Starting point is 00:17:52 there's bullets underneath it that says, what is buying the dip? And the three bullets under it are, one, lower cost buying assets when prices fall from recent high to long-term hope, expecting the price to go back up over time. And then the third bullet is just snack option, which is purchasing dairy or bean dips at the grocery store. This is incredible. Bro, did you buy the dip? There are a lot of possibilities there. It's very good.

Starting point is 00:18:15 It is incredible. This is just a thing that people are checking and seeing. I'm confident this is fixed now, but I've seen this in a bunch of places. Now, you can say maybe it's fake. Maybe you got hoaxed. I've seen answers like this across Google over and over again. And I think these are the experiences people are having with Gem and I in AI overviews. And obviously, Google has to run AI overview.

Starting point is 00:18:38 use as cheaply as possible because they serve millions upon billions of queries every single day. But I think the combination of maybe at the bleeding edge of the frontier, there's a class of people in Silicon Valley who think Google is behind and the next model will rocket them to the front and that race will just keep going. The experiences consumers are having with Google's free AI products and the sort of messiness of how AI is being like smeared all over everything else in Google is a lot. And I would couple that to stories we ran this week about Reddit, Reddit, Reddit being infested with AI Slop to try to game Google search results, Reddit trying to fight back against AI Slop with a moderation of its own.

Starting point is 00:19:16 Publishers wanting to up out of Google because there's no more traffic, which we keep talking about. And you're like, oh, the core business of Google is under a lot of risk from a lot of different places. And if people stop trusting Google is the primary source of information and it's not operating at the frontier and people turn to Anthropic or Open Air or whoever else first, things are pretty shaky, pretty fast. Yes. Right? And like the fact that the company is printing money and doing all the stuff,

Starting point is 00:19:43 it kind of doesn't matter if Google, Google search is shaky. Because once that goes, all of your distribution advantages start to go with it. Yeah. I mean, I agree with that.

Starting point is 00:19:54 And I do want to turn to Reddit, but I just, before we do, I want to offer you one data counterpoint that I've been thinking a lot about this week. So, uh, we've talked before on this show about the fact that open AI,

Starting point is 00:20:06 very proudly sort of touted every milestone on the way to a billion users for chat GPT and has not yet announced that it had a billion users of chat GPT. It did announce that it has a billion people using its models, which is not the same thing, but it's still a really big number. Meanwhile, Google at the end of July, I think, in its earnings said that 950 million people use the Gemini app every month. Like, there is a real chance in there that more people are using Gemini than are using ChatchipT, and if not, it is damn close. Sure. And that is a thing that if I'm open AI is terrifying.

Starting point is 00:20:42 Because that means you're losing, your first mover advantage has now disappeared. And Google is just going to eat you alive by virtue of its overwhelming hugeness in every imaginable way. Yeah, but I'm saying to change the voice in my EV, I had to download the Gemini app and change it in the app. That's fair. Google can play a lot of games that Open AI cannot play. Yeah. I think Google is also in more countries. which I'm an eye than opening eyes.

Starting point is 00:21:08 Yeah. And again, I think Google has a... These are the intrinsic advantages I'm talking about, right? But this is how you get people into the ecosystem. And Google just has moves upon moves that no one else has. And so I kind of think Google can be third best and still win. Yeah. And I also think there is a strong chance that, as keeps happening, Google will find a way to leapfrog the other two for 10 minutes

Starting point is 00:21:31 until somebody leapfrog. Like, this is just what these three companies have done for years. And I don't know of a real thing. reason to assume that Google is like out of this fight. Oh, I don't. There's no way I would say it's out of this fight. All I'm saying is, if you want to lead AI at Google, a good question is, what does that mean? Yes. Does it mean you're in charge of the models? Does it mean you're in charge of how it shows up in Gmail? Does it mean you're in charge of whatever bad voices Gemini has, which are not as good

Starting point is 00:21:56 as the voices Google Assistant has? By the way, Apple's about to put out like a home device with Siri AI and it powered by a distilled version of Gemini that is probably going to be better than the Google home device. Yeah. And I'm only saying that because they will have paid any attention to it at all. Yeah. No, I agree with that. I think Google shipped this home speaker and then I think promptly forgot that it existed,

Starting point is 00:22:20 which is a very Google thing to do. So if you're in charge of AI at Google, like there's an all-powerful robot voice assistant in everyone's kitchen. Well, I think that's what Open AI is going to ship first, too. So are you in charge of that? I do not know the answer to these questions. I don't either. I've been reading a bunch about Google Plus again recently,

Starting point is 00:22:41 as I want to do from time to time. You love reading about Google Plus. I think Google Plus is one of the most fascinating moments in certainly my career covering technology because it was the other time that Google did what it is doing right now, which is take a thing that it believes in and try to smush it into every other product that it made. Social was the future of everything.

Starting point is 00:23:02 They were going to rebuild every single one of their products, around this social network that was not going to just be a social network. It was going to invade every other. Like, what if we make Google Calendar a social network? Was the Google Plus plan? Like, that was literally what they were doing. And it didn't work. And it cratered spectacularly.

Starting point is 00:23:18 And it hurt Google for a really long time. And the bare case for a lot of this is Google repeating exactly that mistake. And instead of saying, we have a bunch of successful products. Maybe this is a new one. They're saying, we need to reinvent every single. thing that we do around this other thing that now exists and they might break it all again in the process. Except, and I think this is the error, it's not all the same thing. If I could reliably count on Gemini to be the same Gemini in every place where I encounter a Google AI service,

Starting point is 00:23:51 I might feel differently about it, but it is often just like nerfed in extremely confusing ways. Yeah, that's such a fun UX question. Like, should you be able to ask Google Maps about your email is such a fun, weird question. There was some announcement this week that you can now like order food directly through Google Maps. You can say, I, you know, I have this long drive. Order me some Chinese food to be picked up along the way. Boy, can I not wait to test that out in a bunch of deeply weird ways that are going to screw me up. Like the Starbucks app can't figure out that I don't live in California anymore. I don't have a lot of faith in Google Maps to get this right. But this is the sort of thing that I think like beyond all of the technology questions,

Starting point is 00:24:32 is sitting in front of a million complicated user experience questions about actually, is the AI supposed to be the same thing in your speaker, in your house, and your email, and everywhere else? I don't know. But it's a super fun, weird question that Google is going to have to answer kind of before anybody else. No, I think everyone else has already answered it, and the answer is yes. Really? Yeah. And Claude is essentially the same cloud everywhere you encounter Claude. And whether that's on the app on your phone or your computer, like these are all cloud services.

Starting point is 00:24:59 That's true. And you can kind of remote-controlled co-work from your phone if it's running your laptop and doing laptop stuff. I guarantee you Apple, we haven't seen how Siri will come to other devices, but they really want Siri to be a unified AI across their devices. And some stuff happens locally and not. And we'll just like figure out where that gets to. But I think everyone else has decided it's the same AI. And Google, for I think a variety of cost reasons, has made the AI on Google Home a different thing than, the AI that shows up in AI overviews,

Starting point is 00:25:31 then Gemini, which shows up in all these other places. Yeah, I think that's right. At some point, that's got to come together, or in classic Google fashion, maybe get even farther apart. Who knows? Well, that, yeah. Or actually, I think the classic Google fashion

Starting point is 00:25:45 would be for them to almost work together, but never actually quite work together. That's the true Google way. It's like, we all acknowledge that these other products exist. We just don't really do anything about it. Yeah. Before we get off this subject, like let's talk about Reddit just for a minute because you brought Reddit up.

Starting point is 00:26:01 Reddit also reported earnings had some interesting stuff to say about their sort of ongoing case to exist in a world where Google is extremely reliant on Reddit. This whole turn has been really fascinating, right? We've been talking about it a lot. Google has understood that it is being overrun by AI Slop, which is very funny because Google is also contributing to a lot of this AI Slop. AI overviews are a disaster for all the reasons you're talking about. but Google has made several turns over the last couple of years to more emphasize human stuff on the internet.

Starting point is 00:26:36 There's the section of Google search results that's now all like creator-made stuff. Yeah. There has been an increasing focus on Reddit, Miyasado on our team wrote great piece. And she said that according to data from Sem Rush, Reddit was the most cited domain in May by Gemini and AI mode on Google. Like, it is looking for these human things. And yet at the same time, Reddit is turning around and saying, oh, my God, we have a massive AI problem. We are being completely overrun by AI slop, AI generated posts. And importantly, people trying to game these AI systems.

Starting point is 00:27:14 Yeah, particularly brands. Yeah. Right? If you can get yourself onto a list of best skincare products, that's the example of Mia's piece. In some Reddit community, the chances that your product will be referenced. recommended by a Google AI overview or a perplexity AI review or chat computer or whatever goes up and that is a gold mine. Right.

Starting point is 00:27:34 So all of this stuff is being flooded by branded content, like masquerading as real people, which is crazy and I think you can do with AI. So Reddit has to deploy AI to save that off. Google is doing nothing. Reddit doesn't think Google is paying enough money to license all the data on Reddit. And you're just like, why is Google's biggest dependency one platform? Right. This is not actually the web you want.

Starting point is 00:27:59 Google's power over many, many, many years has been able to say to any one web provider, we don't need you. Yeah. Right? And this is the dynamic of you have to be indexed by Google. You can never body up to Google and say, we want terms. We want business terms for you to index us. Google will say, we don't need you anyway. Someone else will show up and we can index them and that's fine.

Starting point is 00:28:21 Or the converse to lift something out of obscurity and be like, this is a good and valuable and useful source. we are going to give it real shine. Yeah. I mean, and this was the power of Google search was that it accessed the whole web. Sundar would always say to me, it's the richness of the web that makes Google go. And if you're like, now we're just in a fight with Reddit.

Starting point is 00:28:41 Our whole business is like, can we find a way to thread the needle so that Reddit is healthy so that AI reviews can surface some combination of YouTube shorts and Reddit? You're like, what has search become? What will Reddit become at the end of that? If it is just brands firing AI slop at communities that don't want it in order to sell more shampoo in AI aggregated summaries on other platforms, all of that seems very unhealthy.

Starting point is 00:29:10 And I don't think any of these companies, Reddit or Google included, are thinking about it holistically. I think they're just trying to get through the next term. They're all caught in the same spiral, too, which is that the more important your platform becomes the more AI stuff gets. gets pointed at it, which makes it worse for actual humans. So the humans leaves, so the AI becomes more powerful. Like, it all just kind of death spirals as soon as you add AI creation into the process. Like, I spent a lot of time on productivity subreddits because this is the particular kind of like disease that I have on the internet. And one of the things that has started happening is there are two kinds of posts you see all the time now.

Starting point is 00:29:53 one is opposed from somebody being like, what's your favorite note-taking app? And then a thousand people all essentially hawking their own app at various levels of subtlety. Some people will bury it in a list of like, here are the five that I really like, and number four, like three are like the three everybody knows.

Starting point is 00:30:14 And then number four is a weird, bizarre one that sucks that there's no way anyone would actually recommend and is clearly being the, or you just get people who will, just paste a link to their thing in and hope that that's enough. And what it turns out is that a lot of this is an effort to game the system so that when Google's crawlers go looking for best productivity app, they find the thing that says, what's the best productivity app, scans the list and fights it back to you on Google's AI overview. And it's like, A, we've just completely poisoned the well

Starting point is 00:30:47 because Google's crawlers believe Reddit is full of humans and Reddit is demonstrably no longer full of humans. And B, this is now a disaster for both of these platforms. Yeah. It has made them both worse to the point where I don't really trust anything I see on Reddit anymore because everybody is either trying to sell me some weird thing that they vibe coded

Starting point is 00:31:07 or is astroturfing for some brand or another all over these platforms. Yes, I agree with that. Let me challenge you. Put an over under on this. Okay. Google buys Reddit. Oh, interesting. I would say no, only because it, having Reddit be independent gives Google better reason to prize Reddit.

Starting point is 00:31:33 Interesting. I think, like, Google goes really far out of its way to not remind you that it owns YouTube. You know what I mean? It's really important to Google that you forget the Google owns YouTube. And in such a way that I think Google buys Reddit. just changes that relationship in a way that is hard for Google. It's also in an ongoing antitrust violation kind of way, it would be very hard for Google to be as kind to Reddit as it currently is if it owned Reddit.

Starting point is 00:32:03 Every case for Google only caring about sending you to other Google sites sort of dies on D, but Google loves Reddit so much. And if Google owns Reddit, that falls apart. would it make Google's results better to own Reddit? I don't actually, I don't think so. All you're doing is making it more obvious how to game the system. All right, let me spin it out for you. I'm not saying I believe this.

Starting point is 00:32:27 I'm saying I can spin it out for you. I'm listening. I'm ready for it. Big publishers are all mad at Google. They all start de-indexing themselves from Google because they don't get any traffic and they don't want to participate in AI or use for free. Google says, no, we're never going to pay you. We're just, you can de-index. You're out of Google.

Starting point is 00:32:43 The amount of information in Google starts to put. plummet. It gets replaced mostly by branded content on the web, right? E-commerce providers start publishing lists of best sunscreen. And the wirecutter is not there because the New York Times is mad at Google and D-endix itself. And so Google looks at the incoming information that it has and says, we don't want to surface a bunch of e-commerce information either. We need our own first-party platform like YouTube that is full of something approximating real. At the same time, Reddit says, our AI costs are skyrocketed. because we're spending more money on AI moderation to battle the influx of AI bots.

Starting point is 00:33:20 And this is getting unsustainable at our current burn rates and advertising rates. We need a buyer. And Google says, screw it. We're done with the web. We're just going to index Reddit. The same way that YouTube is not really an open ecosystem. It is a totally closed ecosystem. It allows you to link out, but it is only really indexable by Google search.

Starting point is 00:33:37 Okay. That's a good theory. I think a Google pays Reddit. What, $65 million a year? I think it's $60. 60. Okay, that's probably... You threw in an extra five, like a Steve Huffman shut up. Exactly.

Starting point is 00:33:49 I think that's probably a better deal than buying Reddit outright. And B... Well, Reddit thinks that's not nearly enough funny. True, but what's Reddit going to do? You know what I mean? Like, Reddit has kind of played itself here in the same way. What if Reddit de-ndexed itself in Google? Reddit dies.

Starting point is 00:34:06 This is the problem. This is the place everybody has landed is Reddit failed to make itself enough of a destination so that it doesn't have to rely on Google for people to find things. Have you ever tried to search something on Reddit? It sucks, dude. Like Reddit answers, I don't even know if Reddit answers is still a thing. It's bad. They tried to do their own AI search. And now they're like, well, Google is our discovery engine. And it's like, buddy, have you seen the media industry for the last 20 years? Like, we tried this. It didn't work. But now they're sort of existentially reliant on Google, which makes the fact that all of the data coming from Reddit is less and less useful,

Starting point is 00:34:41 a problem for Reddit as well, because it makes it easier for Google. to be like, well, this is not giving us what we wanted. Why would we continue to license this from you at huge money? They're stuck in this weird mutually assured destruction, but it's all dying anyway sort of moment that just strikes me as very complicated. But the other thing I want to say, and then we should actually take a break here because we have other stuff to talk about, is that there is going to be someone at Google who is listening to this and thinking,

Starting point is 00:35:06 God, what a great case, Neelai just made for why we should resurrect Google Plus. And I want to tell you, it's always Google Plus. Don't do that. Don't resurrect Google Plus or Google Buzz or Google Wave or any of the other times you have kind of sort of tried to do this thing. Nelai's right, but you Google can't do it. So stop trying. I just want to correct you lightly.

Starting point is 00:35:29 One, not on Google Plus. Don't resurrect Google Plus or Buzz. Google Buzz riffs. We're being a little apocalyptic here because I think we see the end game of all this circular bot activity and the rise of the dead internet. right now these companies are making money. So Reddit just had earnings five, six days ago. They beat the street. They posted revenue of $805 million.

Starting point is 00:35:52 Their daily actives are up 18% year every year. Their weekly actives are up 24% year over year. This is a platform that is growing, right? Yes. And it's growing because of a massive influx of Google traffic. Yes. Because Google is existentially dependent on Reddit. And Reddit feels like it's existentially.

Starting point is 00:36:11 depending on Google. And all we're spinning out is not calamity today. It's, boy, if Reddit is overpolluted by AI and Google searches us overpolluted by weird AI overviews sourced by AI and Reddit, like actually this is all pretty precarious. It is pure dead internet theory stuff. You get you get bots that you get bots on Reddit that make bot stuff on Google happen. The Google stuff sends you back to the bots. The bots have more incentive to post bot things that go back to Google. Like, it is just, it's a death spiral of AI everything all the way down. But right now, it works.

Starting point is 00:36:46 There are a lot of corners of Reddit that are still really great. But there's a lot of stats out there about how much of both Google search and Reddit activity is commerce based. And boy, can I tell you that everybody trying to do anything in the AI world is going to desperately try to destroy that with optimization and bots. It is just going to happen. We will end here. If you want the most striking counter argument to the crisis in men in this country, just go to R-slash-smoking, which is just men being very kindly supported to each other about smoking meats. That's so nice. Everyone's like, you did such a good job.

Starting point is 00:37:26 Oh, that kind of smoking. Yeah. Oh, okay. Now I'm in favor. I like it. R-slash-smoking. I'm sorry, it's like a lot of people being like, I fucked up a brisket. Everyone's like, it's going to be okay. And it's just very sweet and tender. That's so nice. The meat and the vibes. See, that will not get polluted by AI. No commerce in smoking. Oh, it will. Trust me.

Starting point is 00:37:45 All right, we should take a break. Then we're going to come back. We're going to talk about all the weird stuff happening to Disney Plus. We'll be right back. Support for the show comes from Bitwarden. Quick gut check. How many of your passwords are actually strong and unique? If you had to think about that for more than a second, today's sponsor is for you.

Starting point is 00:38:06 Bitwarden is the trusted open source solution for pass key, password, and secrets management built to make secure password management accessible to everyone. With its free forever plan for individuals, you can create, securely store, and auto-fill strong, unique passwords across all your accounts and devices. So you never have to remember dozens of complicated logins or rely on the same password everywhere. It also makes signing in faster by securely auto-filling pass keys, passwords, payment cards, and identities unverting virtually any device. Behind it all is end-to-end encryption that helps protect your information from fishing attacks and data breaches. Get started with a basic free plan for individuals

Starting point is 00:38:49 or protect your business today with a free trial. Visit bitwarden.com slash vergecast to get started today. Support for Vergecast comes from Anthropic. Complex work problems need tools that are up to the challenge. Thankfully, there's Claude. Claude is the AI for minds that don't stop at good enough. It's the collaborator that actually understands your entire workflow and thinks with you, whether you're debugging code at midnight or strategizing your next business move. Claude extends your thinking to tackle the problems that matter. You could delegate tasks of all sizes, from tedious bug fixes to new feature ideas,

Starting point is 00:39:31 then handle other work while Cloud runs in the background. Cloud has checkpoints that let you roll back to any previous state instantly. Plus, you could set up workflows to run on a schedule, such as Daily Rep. ports, weekly summaries, and reoccurring data pools. For problems worth solving, get started with Claude at cloud. a.ai slash verge. That's clod.aI slash verge. And check out Claude Pro, which includes access to all the features mentioned in today's

Starting point is 00:39:56 episode. clod.aI. slash verge. Support for this show comes from Odu. Running a business is hard enough. So why make it harder with a dozen different apps that don't talk to each other? Introducing O-DU It's the only business software you'll ever need. It's an all-in-one fully integrated platform

Starting point is 00:40:17 that makes your work easier. CRM, accounting, inventory, e-commerce, and more. And the best part, O-DU replaces multiple expensive platforms for a fraction of the cost. That's why over thousands of businesses have made the switch. So why not you? Try O-D-O-4-3 at O-D-O-D-com. That's O-D-O-O-O-O-O-com.

Starting point is 00:40:39 All right, we're back. Nilai, can I interest you in the following app from Disney? One place to manage all of your cruise plans, your theme park rides, your toy purchases, and all your favorite shows and movies from Disney. I'm talking about the Disney Everything app, also sports, also all the stuff on Hulu. What if everything that ever happened that anyone at Disney had ever done in their entire lives all lived in an app and we called it Disney Plus? This appears to be Disney's plan.

Starting point is 00:41:16 What do you think? You know, it's like they watch succession and they think it's a documentary about success, not death, like success. And it's like, no, succession. It's like a... That's what I got. That's my hot take on what is happening of Disney.

Starting point is 00:41:36 Okay. Tell me more. Well, I just feel like the idea that you run a big empire and you should have one interface to that empire, is very tempting to a certain class of suits who don't know how to make anything. Broadly, I think that's how you view these kinds of grand ambitions. And everyone looks at China

Starting point is 00:41:59 and is like, WeChat is the Everything app. And you're like, yeah, but the Chinese government kind of pushes everyone using like an Everything app and it's really a messaging app with all these other components built into it. Disney Plus doesn't have that. You probably don't want a messaging app in Disney Plus.

Starting point is 00:42:14 And you're just like, did you stop and think about how they got there? Or do you think people love Bluey so much they're going to do their banking here? I think they think it's the second thing. I really honestly do. I think a bunch of people, and I include like Elon Musk and the X-Everything app in this, believe that everyone wants to do everything all in one place. And that there is something inherently simple about that. that is good and beautiful and lovely.

Starting point is 00:42:46 And so if we can have one thing that you like, whether it's yelling racist things at strangers online or it's watching Bluey, then we have an obligation to put as many other things that you do in your life right next to that thing because it makes it easier and better for you to do. I think that is just straightforwardly backwards. I have been yelling about how everything apps are about

Starting point is 00:43:14 idea for a very long time. WeChat is a product, like you said, of the Chinese government. It's also a vastly different economy and culture and way of interacting with each other. Like, the only way for an everything app to work in the U.S. is for every single thing about the U.S. to be different. Like, go back 50 years and change a whole bunch of things, and that's how we do this. But this, there is something about this idea that is so tantalizing and so sort of explainable. And, which I really enjoy, right? Like you can imagine the meeting at Disney where they're like,

Starting point is 00:43:48 well, we have theme parks. And the people who go to our theme parks, like our movies, what if when you were watching a movie, you could buy a ticket to the theme park? Like, I swear to God, that's how that meeting went. I wasn't there.

Starting point is 00:44:00 I don't know, but I am certain that's how that meeting goes. Yeah, we want to put offers here. And then if we're going to have offers, you might as well complete the transaction in the app. And now we're just going to add a ticket screen and then you've got the wristbands at Disney. You got to manage those.

Starting point is 00:44:11 You might as well manage those in Disney Plus. And we just have one app on the phone. We don't have to deal with trying to get anyone to download new app. And certainly, we don't have to deal with the cost of maintaining many apps, let alone that all these different screens inside these apps will actually just be cost centers on their own, but it's one app and it's self-containable. I get it. Somebody yells efficiency and it's like meeting over. So I guarantee you the thing that scares all of these companies, maybe Disney most of all, and I have some evidence for this, is the social platforms.

Starting point is 00:44:40 is I keep saying if you run a Hollywood studio, you have to contend with the fact that your real competition is TikTok and YouTube and Instagram, and those platforms don't have to pay for content. There's just an army of teenagers who work for free, and you just have to stare at them. You have to think about them every day. And that means that TikTok and YouTube and Instagram

Starting point is 00:45:03 always have something new to deliver somebody, and that means they win a spot on your home screen, because if you're bored in line, You can just push one of those buttons and something will happen. Right. Some new thing, some new rage bait controversy will appear at you. Sure. And TikTok didn't pay a dime for it.

Starting point is 00:45:22 If you're Disney, you're like, we got to make another MCU show. And like, that's a disaster. Yeah. Like it just. It means you have to try hard. You have to do good, expensive things over and over and over. So now you're like, how do I beat Instagram and doing it with the cost model of Hollywood content? is just fundamentally upside down.

Starting point is 00:45:43 How do I get you to open Disney Plus at the same rate as Instagram, but I have to pay Robert Downey Jr. every time he sneezes and TikTok doesn't have to pay anybody anything. You can't win that fight. You will never have enough content. So you can do things like

Starting point is 00:45:55 you can explore adding a free tier to Disney Plus, which is something Disney announced. You can make it the app that everyone has to use. That's going to jack up your DAUs because everyone in the park is going to be using your app all the time.

Starting point is 00:46:06 You can put Disney Cruises deals in it so you can send notifications. So the sorts of people who like to go on cruises, whoever they are, I'm just not one of them. But they appear to be many, many, many millions of people who love to. In fact. Yeah. And then lastly, this is my evidence. You can try to split the difference.

Starting point is 00:46:24 And you can go to TikTok creators and give them official access to Marvel, Star Wars, and Pixar characters to make TikToks with that you then license back into the Disney Plus app so that you have free content in your app made by other people using your IP. And maybe you're just going to. click the Disney button on your phone and you're going to watch some Star Wars cosplay nonsense. This is desperation. Yes. So, okay, you know what you're making me think. This is what you might call the time spent epidemic. We spent a long time railing against engagement as the main metric, right?

Starting point is 00:47:02 The thing that you needed was to make people do something. That if people clicked a button or responded in some way, that was a, success, right? This drove social media into what it is now that if you can make a lot of people angry such that they comment, that post was a victory and you've done a good job and we should show it to more people.

Starting point is 00:47:21 And I think you can make a real case and I have made it before that engagement kind of broke the internet. And we have moved on, or at least all these entertainment companies, have moved on to time spent. Time spent is the thing they all talk about. And it is the race. It's why Reed Hastings said they compete with sleep.

Starting point is 00:47:37 It's why YouTube is competing with video games. It's why everybody is trying to be everything all the time. The goal is just to get you to look at their app more often for more time and what it takes to get there is utterly irrelevant. This is just the place we are.

Starting point is 00:47:53 Anything that looks like time spent wins. Fans will enjoy a shared Disney experience with TikTok videos that will be refreshed regularly to engage subscribers more deeply in timely fan conversations, build stronger communities, and ultimately help them engage with more story formats.

Starting point is 00:48:09 across Disney Plus. It's in the press release. The thing you're describing is nakedly in the press release. We're trying to increase time spent in the app. This is also Facebook testing the thing where it's just vertical video when you open the app because everybody knows that when you show people a video, they stick around longer, they scroll more. Like all of the, it's just time spent. God, that sucks.

Starting point is 00:48:30 I will say, we'll see how real this is. There's a nice part of this where they're going to run a, the Disney Creator Ambassador program. And if you opt into this partnership with TikTok and Disney, you might find yourself a creative who gets a job at Disney because they're going to do career development for creators. Great. Maybe this is the solution to your AI apocalypse of entry-level jobs.

Starting point is 00:48:55 You have to do your own free animation for Disney, and then they will pluck you from obscurity, and then you will run Pixar one day. Mr. Beast, Imagineer. The power of short form video. thoughtfully curated to increase time spent. I'm telling you, all of these platforms are terrified of this, and they're all going to launch stuff like this.

Starting point is 00:49:16 The Everything app, you know, approach to whatever, you can buy tickets here is just one way to increase engagement and increase transactions inside the app. Putting fan content licensed with TikTok officially inside the app is another way to do it. And all of them are chasing the economics of the social platforms themselves. TikTok does not have to do this. They just showed up.

Starting point is 00:49:38 Right. And in fact, maybe the most important thing that TikTok and Instagram and YouTube and all the rest provide is a suite of music licensing deals without which they cannot operate. And that's really the only money they spend out in the world. And the labels know it. And every night and again, the label's bent them over a barrel to extract more concessions. But AI music is coming for that too. And you're just like, oh, this is a big spiral. Right.

Starting point is 00:49:59 We're all chasing the economics of they don't have to pay for anything. Yeah. Well, and I think the other trend we've seen a lot of during this season of earnings is ads. I mean, this has been the trend for a couple of years now. All of these companies, instead of jacking up subscription prices, or in addition to jacking up subscription prices, are desperately trying to get people to switch to plans with ads because they make more money on users who watch stuff with ads

Starting point is 00:50:24 than from your subscription. It's not exactly like that, but it is often like that. And you just put those things next to each other, and there has got to be this belief running everywhere around this industry now, that if we can just get you to spend 90 minutes a day in our app, we will inevitably make tons and tons of money. Yeah. That there is enough demand for these ads and enough supply for them that if we can get you

Starting point is 00:50:47 to open our app for 60 minutes every day, which is like the Instagram, TikTok level of engagement and time spent, we will suddenly just absolutely open the money funnel. And how we get you there, not important. Right. And it's going to be licensing the content from TikTok. Yeah, which is. TikTok doesn't have to pay the money. Oh, God.

Starting point is 00:51:09 And this means we're like one small turn from TikTok videos being turned into Disney World Rides, which is just like not a world I'm ready to live in. I'm ready to live in that world. Like some obscure, extremely dirty rap bridge showing up as a Disney ride. Let's do it. Yeah, meme rides is not a, not a thing I'm fully prepared for. What do you make of this free Disney Plus thing? This is being asked of every streaming service, right?

Starting point is 00:51:35 To be just reported it's growing up. like crazy. And now every streaming service is being asked, when are you just going to go free and ad-supported as a giant growth mechanism? Disney, Josh DeMorrow, the CEO of Disney, said they don't have anything to announce in the way that you say it when you're going to announce it in five minutes. So, like, I'm confident Disney is at least down the road of figuring out whether this makes sense. Is this the next big thing? If you're chasing engagement and you're openly announcing deals to increase engagement, then the last thing you can do is make the thing free to get people in the door.

Starting point is 00:52:09 And I think Disney's initial thesis was where Disney and everyone pays. You don't go to Disney World for free. You pay a lot of money. More all the time. And yeah, an ever-increasing amount of money. And I think they're just up against the reality of, well, the entire Spider-Man movie leaked onto TikTok for free. Yeah. It made a lot of money.

Starting point is 00:52:30 A tremendous amount of money. but the reality is no one's opening Disney Plus for it because they can just steal it. And their subscription fatigue is real. And if they're paying us money and they're not opening us every day, and if the way to get them to open us every day is to have TikTok videos in here, then the money is not worth it. Right. We're just going to shut this down. And we'll just see what we'll see.

Starting point is 00:52:53 But I think the incentives are aligning for a free tier full of users generated content and a paid tier full of Disney content. And if they're smart, they will do that and leave the ticketing experience to some other app. So like Disney Talk essentially becomes free as a curated reason to open Disney Plus, but then if you want the shows and movies, you have to pay for it. Yeah, all that shoulder content, you know, like behind the scenes of the new Marvel show, like make that free in the Disney app, what people have at it. Yeah.

Starting point is 00:53:23 But when you want to click over and actually watch the movie, make that the subscription fee. And no one's really tried that kind of approach. Yeah. But again, if you're desperate to get people to opening app every day, you are going to have to make some stuff free just to increase the number of people who have that app on their phone. Totally. I think that's right. And I think it's interesting to see all of this happening next to the biggest box office weekend in history at the movies. You have this audience telling you they are desperate for like tent pole reasons to go out into the world.

Starting point is 00:53:55 And it's like, oh, actually, if you make a kick-ass movie, lots and lots of people will go see it. which five years ago nobody thought was true anymore, right? Like we were declaring the death of the movie theaters and the box office and in-person experiences in general not that long ago. And I think you can trace a lot of these decisions these companies are making kind of back to that. But there is this resurgence in people wanting to have big, exciting, high-budget shared experiences right next to this just clear race to the bottom from all of the streaming services. and it'll be interesting to see if

Starting point is 00:54:29 anybody can pick themselves out up of that, which is an accidentally great segue into what is going on with Paramount. Oh, God. Because one of the... So the thing that's happening right now is Paramount is trying to buy Warner Brothers Discovery in an effort to, as Nilai keeps saying,

Starting point is 00:54:48 tank the company as everyone does who tries to buy Time Warner or anything like it. But David Ellison, the CEO of Paramount, has been running around to everyone who will listen throughout this whole process saying, my goal is to make bigger, better, more expensive stuff.

Starting point is 00:55:05 We're expanding the slate of movies that we're making. We're going to not join this race to the bottom where everybody's trying to be TikTok and YouTube, but we are going to be and save Hollywood all by ourselves. And overwhelmingly, it seems, every time any U.S. regulator or judge is confronted with this, they say, no, thank you. We are not interested in this.

Starting point is 00:55:26 The latest news was that basically this deal has been halted. It's going to go to trial over some of the questions about the antitrust implications of all of this. And that trial has been set to begin next March, which is a long time from now. And it's a particularly long time if you are paramount. And what you agreed to do is start paying what's called a ticking fee if the deal doesn't close by September 30th, which I will remind you is well before March of 2020. Seven. So this is earlier than I think they wanted. They had proposed to kick it down the road and settle, basically.

Starting point is 00:56:05 Yes. And I think the judge said, no, we're just having a trial. So now you've got to prepare for trial. This is like you're not negotiating against a bunch of state attorneys general. You're preparing for a trial that they are all pretty motivated to win because you are the boogeyman. Yes. I don't think Ellison is wrong in saying a lot of this is about concern over who will run CNN and CBS News. I think that he-

Starting point is 00:56:25 Do you want to read the quote? You should read the quote. It's a good quote. So David Ellison wrote an op-ed in the New York Times that I think the essential implied title is like, I'm cool. Yeah. It's actually, it says David Ellison in defense of the Paramount Warner Deal, but it is essentially a defense of David Ellison by David Ellison. You can trust me. Yeah.

Starting point is 00:56:46 Here's two quotes from this. This is Ellison and the Times. I believe this fight is not really about market share. If it were, it wouldn't have been reviewed and approved by regulators reflecting 65 countries, including the United States of China, as well the European Union. I point out that the United States federal government reviewed this deal and was essentially bribed by Paramount. So whatever.

Starting point is 00:57:06 I believe a plainer where he sits beneath the briefs and news releases the news. The issue is where I can be trusted as steward of Warner CNN. There's some speculation about my politics. My lawyers lose my attention. Unfortunately, I can't give anyone a view into my heart and mind. But I can share this. I've regularly voted for candidates for other parties. I hold some views that we call conservative and others who we call liberal.

Starting point is 00:57:26 And when it comes to our news operations, I do not aspire to leave these companies to bend their newsrooms from my views. I believe that news should be based on facts and truth. Which is what everyone says before they bend the news to their opinions. I just want to point out that we can see what happened to CBS News. It's not like a secret that in order to get approval to land that deal, he agreed to install Barry Weiss as the editor-in-chief of CBS News, and she has torn it to the ground. many, many allegations about Barry Weiss and her regime tearing CBS views the ground, including at 60 minutes, including political interference into shows. And the ratings are in the toilet. Like, the ratings have just gone down.

Starting point is 00:58:04 And I think it's because the viewers of broadcast news, particularly CBS are older, and they expect a certain level of quality from CBS, not just craziness. And, like, outright in order to get a deal, we brought the Trump administration by installing this figurehead on top the news. Like, that's the thing that happened. It's the thing that there's reporting that Ellison has said to the Trump administration about CNN in order to get approval to buy Warner. So this is all great, right, that you can publish this thing. You say you're going to leave it alone. But the evidence is that you will not leave it alone. Right.

Starting point is 00:58:37 That in order to win continued approval or to do more deals, you will actually continue to screw with things. And maybe this doesn't matter because young people are not watching cable news anyway. They're certainly not watching broadcast news. it will come to in the Brennan Carr segment. But, man, you're just like, guy, you're up against YouTube. And the only part of this where he even gestures at it is by pointing out that when he buys Warner, he will still be pretty small. So here's another quote from Allison in this piece.

Starting point is 00:59:05 Add up every hour American spend watching television and a combined Paramount Warner would account for less than 20% of all watch time competing every day against Netflix, Amazon, and Apple, all of which are companies whose resources dwarf hours, include YouTube's user in area content math and our share would drop to 13%. A combined company would account for just 18% domestic box office. These are not the numbers of a company that can dictate what audiences watch or what writers get paid. That's so bonkers. Those are such giant numbers. They're huge numbers. To say out loud, my company accounts for 20% of all watch time in America and be like, but that's nothing, right? It's like crazy. Right. And you're like, yeah, I understand

Starting point is 00:59:46 what you're trying to say, but these are huge numbers. These are, you're going to run a big company that's in charge of a massive IP, and you're going to run CNN, which you are admitting is the thing people are worried about. And then you're trying to convince us that your company won't have huge cultural impact. I would also point out that Oracle, which is run by Larry Ellison, David's father, owns 15% of TikTok US and operates it. And effectively bankrolled when Skydance bought Paramount and put David Ellison in his position in the first place. And is bankrolling this?

Starting point is 01:00:16 And is bankrolling this? Yeah. So you have a massive cultural influence operation just running out of the Ellison family. And obviously structurally, the corporate structure is more diffuse. But it's the Ellison family owns and operates TikTok, owns and operates Paramount, and maybe we'll operate Warner. And you're like, yeah, I think these numbers, I think you're big, there's a reason you pick YouTube here and not TikTok. Right. Right.

Starting point is 01:00:41 Like, this is a lot of cultural influence. I think a huge problem for all of this is that Oracle's stock is down 26% year to date so far. Right? The pressure on AI stock is coming. There's a hedge fund this week that almost had a margin call situation because it was all in an AI. Whether or not there's enough money in the bank to fund ticking fees and the restructuring of all these operations while Oracle has to contend with the reality of the AI trade, I think is the biggest open question of all. And I think there's some people who are like, maybe this deal isn't going to happen now. Yeah, there is an increasing sense that the sort of inevitable outcome here is that Paramount just walks away.

Starting point is 01:01:24 Because as you're saying, it would add up to if this whole thing goes through in next June, it would be, I think, somewhere between $1.5 and $2 billion of ticking fees. Just ticking fees. And this for a company that has already, I think, been forced by Netflix to massively overpay for Warner Brothers Discovery. Larry Ellison can swing it. Like, dude has enough money. But you wonder at some point when he's going to feel like he's just throwing bad money after good, potentially given the fact that they might lose anyway. Like, they might, we're having midterm elections in a couple of months before this trial. They might just lose anyway.

Starting point is 01:02:04 And if the AI trade continues to go badly, Oracle stock continues to go down, data center protests continue to go up, you're not throwing away money that you're going to make back just because AI is going to boom. Totally. And who, like, I don't know. Maybe it's all going to work out for Oracle. Larry Wilson has survived worse. Yeah. But the pressure on this deal is coming from a number of different directions. And I would point out, you're operating TikTok.

Starting point is 01:02:29 So what you really want is cultural influence, you can go screw with those knobs a little bit. Yeah. Yeah. All right, while we're talking about TV news, we should obviously talk about Brennan Carr. So let's take a break, and then we're going to come back, and it's lightning round time. We'll be right back. Support for the show comes from KPMG. In any organization, disruption is inevitable, but struggling through it doesn't have to be. The KPMG Adaptability Index is your blueprint for building capabilities to handle what comes next. It uses real data to look at how your culture, strategy, and partnerships all work together to help your business thrive. Stop reacting and start adapting.

Starting point is 01:03:11 Visit KPMG.com slash US slash adaptability to explore the Adaptability Index and Pulse Surveys today. Hey, it's Ryan Reynolds here for MintMobil. Now, I was looking for fun ways to tell you that Mint's offer of unlimited premium wireless for $15 a month is back. So I thought it would be fun if we made $15 bills. But it turns out that's very illegal. So there goes my big idea for the commercial. Give it a try at mintmobile.com slash switch. Up front payment of $45 for three months, $90 for six months, or $180 for 12-month plan required.

Starting point is 01:03:47 $15 per month equivalent. Taxes and fees extra. Initial plan term only greater than 50 gigabytes may slow when network is busy. See terms. Do you love the luxuriant flavors of fine dining? Then make a reservation at the most hands-off restaurant in town. Udicef Five-star chefs at U-Dishaf

Starting point is 01:04:03 provide expert suggestions on how to prepare your meal so you pay them and then they say au revoir while you try and figure out the difference between Esca Go and Escarole No dish is too simple or too fancy for us at U-Dicef because there are literally no cooks in our kitchen

Starting point is 01:04:18 Just you all alone Prepping produce you don't even know how to pronounce Bon Appetit Fortunately You de Chef isn't real That would be ridiculous. Kind of like AI that suggests what your employees should do, but can't help them do it. That's like a restaurant that only tells you how to cook your own meal.

Starting point is 01:04:39 ServiceNow doesn't stop its suggestions. It puts AI to work, connecting your AI, workflows, data, systems, and governance to take action and get the job done. That's why the world works with Service Now. All right, we're back. It's time for the lightning round, which means it is time once again. for America's favorite podcast within a podcast, Brendan Carr's a dummy. All right, now it is time for America's favorite podcast within a podcast.

Starting point is 01:05:09 That's good. What did Brendan do this week? Let's start with the basics. Brennan was particularly dumb in two different ways. Is it dumb to take bribes as a public official? Generally, yes. I feel like Brendan's knights consist of coming to defend any else. We still do.

Starting point is 01:05:32 I think you can. Classic dumb brown interests you never get it. This is all so much more than Brendan deserves. Because Brandon. It's good. First of all, never play my voice at me. That's just rude and I don't appreciate it. It was good.

Starting point is 01:05:53 I enjoyed that. It's a slow vibe for today's Brendan, which I enjoy. That came from Brady Garaki, who says they did not use AI. Thank you, Brady. Appreciate it. Neli, what did Brendan do this week? We got two things this week. One is straightforward FCC news.

Starting point is 01:06:10 They did stuff. They took a vote. They passed a thing. We can talk about it. And then the second is what is fast becoming, my favorite. favorite theme on Brennan Carr's dummy, Neely agrees with hardcore Republicans. The podcast within a podcast within a podcast.

Starting point is 01:06:23 Yeah. In which Nealai, the Wall Street Journal It's called strange bedfellers. Extremely hardcore Republicans come to the same conclusion as me, which I think is truly a delight. So this week, literally, just before we started recording, the FCC, which is supposed to have five members, but only has three per project 2025,

Starting point is 01:06:43 voted to reduce the broadcast. limit. So you're previously only able to own 39% of stations in a market. Brendan has been fulminating about how this is bad for ages and they were going to do this vote and they did the vote. And now it's a case by case basis. So if you want to own more than 39% of broadcast stations in a market, you can go to the FCC and you can make your case why it's good or bad, there's a lot of legal wrangling to come, whether or not this is actually legal for the FCC to do is a thing FCC commissioners and many administrations have said they have the power to do, but which Congress has said, no, you don't.

Starting point is 01:07:17 Lawsuits will ensue. Brendan's just kind of been doing it, though. Like, we've been talking about the Sinclair Technomerger for some time now, right? That, like, this is... He's just been doing it. Yeah. He's just been doing it. And people, again, the lawsuits are to come.

Starting point is 01:07:30 Brendan's arguments for this are particularly dumb. Sure. So he was on Fox business earlier this week, talking to Marina Bart Romoomo about the vote. You know, it's a particularly good setup because he's a particularly good setup because he's He's like, I don't know how we're going to vote. We're going to look at it. Ever knew how he was going to vote. So here's Brendan before the vote explaining why he wanted to lift the count.

Starting point is 01:07:53 We've been way out of balance in broadcast media recently. You've got these what we call national programmers like Disney and Comcast, and they produce basically all the shows that people see, but there's thousands of these individual TV stations out there that want to serve their local community. But the national programmers, Disney and Comcasts have gotten so powerful that they just bowl over these local newsrooms. So by lifting this cap, you get some scale on the TV station side so they can finally push back and give some balance to the national programmers. That's going to be more local news, more local reporting, and that's more trusted than the national media.

Starting point is 01:08:28 So I think that's going to be a good thing. All right. Let's just try to pull that apart. The national programmers, like Disney and Comcast. By the way, the fact that he just said that as if he just invented the phrase national programmer, insane. Sorry, keep going. But the idea that Disney and Comcast are bowling over the locals is very funny because they only produce one national news broadcast on their networks. Right.

Starting point is 01:08:57 Like, they're not in the local newsrooms of ABC 7 in Chicago. Do you know what I mean? Like, that is just not how that works. And I'm sure we're going to get a bunch of people who work at local stations, emailing us to explain exactly how works. But like in the broad strokes, the local franchises do what they want. They produce their own news and they put it out in the world. Comcast, by the way, also kind of desperately trying to get out of the content business because this all turned out to be a terrible decision for Comcast.

Starting point is 01:09:25 So very funny to make Comcast the like big bad in this. I mean, none of this makes sense just from the mechanics of it. But in like classic Brendan fashion, right? You're like you're conflating a bunch of ideas and talking really fast in order to say something stupid but make it sound smart. Brennan Carr, everybody. It sounds like podcasting. So the idea that the national broadcasters, the national programmers are forcing stuff down the throats of these local franchises just kind of broken if all you're talking about is the news.

Starting point is 01:09:56 Right? Like that's just not what's happening there. It might be that maybe the local stations don't want to run all of the sitcoms that NBC produces. But like, fine. Do you know what I mean? Yeah. That's not what he's talking about. He's talking about trust in the news.

Starting point is 01:10:12 So this is already like, we're already confused about what the stations have to run, what they hold themselves out at, and who makes the news. And then there's this claim, the heart of his argument here. By lifting this cap, you get some scale on the TV station side. That's going to mean more local news. I'm sorry. Do you know who's buying the TV stations? It's giant private equity firms and other national broadcasters. And they're not hiring more reporters.

Starting point is 01:10:38 They're not making more local news. The economics of putting more reporters in Kansas City and Milwaukee and Chicago and Detroit are not served by additional scale. Do you know what I mean? It's the opposite. It is in fact. In fact, chasing scale is what killed all of this stuff in the first place. Right. Because, again, Google doesn't have to pay for content and Instagram doesn't pay for content.

Starting point is 01:11:02 So they all have this. And Brendan is constantly saying, look, the cable programmers reach 100% of the country. the internet platforms reach 100% of the country. These broadcasters, if we let them get really, really big, they'll reach 100% in the country too. And it's like, sort of, but there's no connection between that and more local reporting.

Starting point is 01:11:21 Right. You haven't said, if you want to get really big, we, the government, are going to require you to hire local reporters in your markets. Right? And maybe you probably don't want the government to do that anyway. So, and who would watch it? Because who's watching local TV?

Starting point is 01:11:35 It's not young people. watching Instagram. This, by the way, from the same government that is also desperately trying to kill publicly funded media. So I'm confident to say that's not the plan. And then the real answer is who's buying this are Sinclair and Tegna, which are aligned to the Trump administration, which there's endless videos of Sinclair stations in particular forcing their broadcasters to read coverage that is favorable to the Trump administration. You can go search that clip from Deadspin from 100 years ago. They're still doing it. Yeah. What are you talking about, Brendan? Because there's no.

Starting point is 01:12:06 logical connection between Comcast makes you run rerones of friends. And if I let the Sinclair's the world get even bigger, they're going to hire more reporters. And that will increase trust in local news. Like, it is really difficult to draw those logical connections. That is the fundamental basis of his argument. The FCC is obviously supposed to have five commissioners. It only has three. The one Democrat Anna Gomez voted against this.

Starting point is 01:12:31 Of course, she dissented. So you've basically got a situation where, people have to file lawsuits against the FCC's authority to do this. In the meantime, these deals are going to push through at lightning speed. And I think they will result in more consolidation in local media and eventually even less trust in local media. Because if everyone knows what they're listening to is local news brought to you by some giant national conglomeration of stations that is barely paying attention to your small market and probably going to start turning out AI slop across your markets, you're going to go elsewhere. And you're going to continue to turn to actual people in your community who are using tools like TikTok and YouTube and Instagram, thus hastening the death spiral. Yeah.

Starting point is 01:13:16 Hastening is the right word. This has been happening. And what he is proposing is to speed it up. By putting more power in the hands of Trump allies. Yeah. Anyway, that's broadcast consolidation. That's the actual thing Brendan this week. The other piece of this, which I think is just a delight in a recurring theme here on Brennan Corseudmi is the FCC's general counsel was testifying in front of a congressional committee this week.

Starting point is 01:13:36 And Senator John Kennedy, who is, I would say, a very hard right Republican, straight up said, what are you doing, man? We have a clip. Let me just play the clip. Sometimes the FCC scares me right now. We haven't given the FCC the right to regulate freedom of speech, have we? You have required the FCC to regulate broadcasters pursuant to the public interest. And who decides what the public interest is?

Starting point is 01:14:04 It is the judgment of the FCC following precedent and following the law. What if the FCC is in the hands of someone whose politics you disagree with? You're getting into the foothills of violating First Amendment. The foothills do. A lot of bad foothills going around right now. Everyone's in the foothills, just scrambling around. No one knows what happens next. By the way, who decides what's in the public interest is a great question that you can

Starting point is 01:14:34 consistently ask a Brendan Carr because right now it's him. And a thing that he consistently said is the public interest is listening to everything Donald Trump says, which you don't, again, as an American, you simply do not have to, you don't listen to the president at all. It's the best part of being American. Yep. You put a don't tregg me flag on the back of your pickup truck. You'd jump over little cars.

Starting point is 01:14:51 This is just what I dream about at night. It's bad when you have both sides of the aisle. In particular, some of the most libertarian-leaning conservatives saying, hey, you're getting a little, it's getting a little dicey. Yeah. It's like you said, most people agree about free speech, it turns out. We're pretty much all on the same side. Government speech regulations are bad is the real thing

Starting point is 01:15:13 that I repeat all the time. Maybe the only consistent political opinion I've had in 15 years of the verge cast is government speech regulations are bad. And here, it's you know, it's me, it's the editorial pages of the Wall Street Journal and it's apparently center. John Kennedy, who is now invited on to Brennan Carr's

Starting point is 01:15:29 a dummy, America's favorite podcast for the podcast, only to talk about this. Yeah, just this one just this. It's going to be a really short segment. Is you a dummy or not, John? Answer the question. Anyway, Brandon, if you want to come on this or Dakota or really anywhere else at all, you're always welcome.

Starting point is 01:15:43 I would love to hear the mechanics of how lifting the broadcast cap will increase investment in local reporters because I haven't heard it yet. You just keep saying it like it's true. But I think you're a dummy. Anyway, that's been Brandon Carr's dummy, America's favorite podcast. We're in a podcast. Feels good. That's good.

Starting point is 01:16:08 My headphones are a little louder than usual today and really rock the dome. If you've got an EQ, I encourage you to turn up the bass as far as you can when that comes on. It's powerful. It does some things to you. I have a lot to tell you about here in the long run. But we should start with the end of a burgeoning segment here on the Verge cast. It ended so fast. I know.

Starting point is 01:16:31 It really, I think we only got through two Nikita Beers going through. moments here on the Vergecast. Nikita Beer, who has been for a while the product chief at X, posted that he is leaving X. He wrote that, ladies and gentlemen, it's time to pass the torch and demote myself to my natural state, a poster. This is all, I would say, not terribly surprising. Nikita Beer has spent the last several months posting on X about how unpleasant it is to be running product at X. He's been publicly running a series of experiments, all of which have come to the same conclusion, which is that X is bad, and so has everyone on it, And if we take, literally, like, if you put all of his studies together,

Starting point is 01:17:08 it says that if we take all of the people and posts off of X, everyone likes it better. It's like, this is the state of X as a platform. So he's leaving. He's out. They shipped a bunch of new products. He's been there for, like, a little over a year. It's got a real, like, defense against the dark arts professor vibe.

Starting point is 01:17:29 Like, there will be millions of people who try to do this job, and none of them will do it successfully. But I just want to say, you know, Nikita goodbye. You, I don't know if you did your best, but you did something. You did it. In the grand scheme of things, he is the most successful product chief that X or Twitter has ever had. For sure.

Starting point is 01:17:47 He shipped a lot of stuff. Including a robot that published CSAM all by itself. That wasn't in his little list of things that X had published. No, sure wasn't. Grock calling it Mecha Hitler is not in that carousel, as far as I can tell. a nudification problem that threatened to bring the whole thing down, didn't make his list. An algorithm that openly favors Elon, whenever Elon wants.

Starting point is 01:18:13 Elon openly overruling the keyest decisions left and right by just posting, we'll look into this. It didn't make it onto his list. Yeah. That platform is in some kind of decline. We know this, actually. This is my lightning round item because SpaceX had its first earnings after the IPO. Liz wrote it up.

Starting point is 01:18:31 People are very mad at Liz because they don't. People love SpaceX. People love rockets. The problem is that SpaceX now encompasses many of Elon's other companies, including XAI and X, uh, platform, which is failing. And the product just left after again, overseeing an AI trap bot that referred to itself as Mecca Hitler, denutified celebrities and women at will, and, uh, produce ESAM. Tight.

Starting point is 01:18:56 Yes, the rocket part of it is cool. The rest of the business. in disarray. Liz points out in her piece, her analysis of the earnings, that the business of SpaceX is barely space and mostly XAI. XAI has a ton of data center commitments

Starting point is 01:19:13 that are holding company underwater. And then importantly, and I think we're going to see this really play out, the only truly profitable part of SpaceX is Starlink, which is a telecom. And Gwen Shotwell on their earnings call said they have ambitions to come after

Starting point is 01:19:30 terrestrial cell networks. So they want to compete with AT&T and Verizon and T-Mobile, which good luck. And I don't say this because AT&T and Verizon and T-Mobile are great companies or well-run or beloved by their customers. I say this because the economics of running a telecom are extraordinarily well-known.

Starting point is 01:19:47 Yeah. And all of those companies struggle in different ways all the time because the build-outs are expensive, the physical plants are expensive. You have to maintain them. You have to uphold them. They hate competition. And so if you come in to compete,

Starting point is 01:19:59 you need lower prices, but you've got the same fixed costs. They all spend billions upon billions of dollars to lobby the government to get sweetheart contracts. You are in for a knife fight. And if what you've got is your phone can connect to space, you're going to run into the problem of trees and buildings quite directly, which means you need to build a cell network, which is why there are rumors that Elon will buy Verizon or T-Mobile. Lots of rumors about Verizon this week. A few weeks ago, it was T-Mobile. I'm guessing next week it'll be AT&T.

Starting point is 01:20:27 Hopefully he gets rights to the gray scale Snyder cut if he buys a lot. AT&T. That really, I think, brings together a lot of threads for us on the show. What about Project Gena 5Sys? Can they buy Project Gena 5Sys? I mean, it's available. It's bound to me out there. But SpaceX does have some spectrum that they acquired by yelling basically a lot, and Brendan gave it to them. We'll see if they can build that out. But again, the idea that the most profitable part of SpaceX will have the economics of a telecom is openly hilarious. It is very hard to make the economics of a telecom.

Starting point is 01:21:00 better than the economics of a telecom. I mean, it's pure Elon Musk, right? Like, this is, he would like you to think that Tesla is anything but a car company when, in fact, Tesla is very much a car company. It's a car company. And they're going to do robots soon. Don't worry. Yeah, sure.

Starting point is 01:21:13 And now he would like you to think that SpaceX is a space company when, in fact, it is either a data center supplier or it is a wannabe telecom challenger. You know, it's a lot. But rockets. Like, what if your whole marketing strategy was to launch and re-land rockets on Earth? Just what a deeply bizarre fact. But that's where we are. So the leasing data center space made more money than the rockets.

Starting point is 01:21:45 It also did more spending. So they spent $15.8 billion on AI alone. Space and connectivity was a touch over a billion. So the data center piece is just like they're a neoclop. And we'll see if that works, right? as we discussed with Google, there's a lot of demand from Anthropic and OPAI.

Starting point is 01:22:02 Those are basically your customers. Yeah. The Colossus stuff that they built appears to have been a shockingly great business decision. Yeah. Right, because people want the capacity. Yeah.

Starting point is 01:22:15 SpaceX is renting capacity to Google, Anthropic reflection, and then they bought Cursor. So we'll just see how all this goes on the AI side. X revenue and all that is down. We've seen that happening. But the part,

Starting point is 01:22:27 I just want to continue to stay focused on. is the idea that Starlink will provide internet connectivity to everyone in the world, which is the thing that they say. And I don't know, man. It's not like I love Verizon, but I got Fios. Yeah.

Starting point is 01:22:39 What are you going to offer me? Trenches to get Internet to my house. It's like a real thing that happened. There's no chance. I'm going to take out my fiber connection and replace it with a space connection. Right. I don't know anybody with fiber who would do that.

Starting point is 01:22:52 And I, the ideas are being floated that like all of the Starlink dishes will turn in this. sell repeaters and become a mesh network. It's like maybe. Maybe you can do that. We're doing 5G again. We're literally just doing 5G again. These are all fun ideas that maybe you can do, but then you got to upgrade all of your dishes. You got to replace all the satellites, which keep falling out of the sky.

Starting point is 01:23:14 Meanwhile, Verizon's fiber has been dug. It's just there. They can keep using it. There's a Verizon guy who comes to my house like every three days to try and sell me internet. Like, is SpaceX going to employ that person? Anyway, the way that you compete in telecom historically is you find a way a cost model that lets you drive lower prices to consumers and that usually requires new technology. And you can spin yourself into believing at Starlink is all of those things. The way that's Sprint, when Sprint first came out, remember they had the pin drop sound?

Starting point is 01:23:43 Yeah. It's because they were running a digital network, not an analog one, and they could bring their prices down with higher quality. The Sprint did not succeed in this podcast until just now. Just pointing it up. Like, there have been lots and lots of attempts to do this. Jenna five says another attempt. We're going to use open hardware. We're going to do open RAND.

Starting point is 01:24:02 Remember, we talked about O-Rand for ages on the show. Amazon is going to, with the power of openness, we will run a cheaper cost model than Verizon, said Charlie Ergan, who just declared bankruptcy. Good luck, man. Like, the economics of telecoms are extraordinarily well-known. If you can do, if he can pull it off, he can pull it off. But I, the number of people in our comments are like, you just hate SpaceX. Like, you know, the Rockets part is really cool. The part where you're going to run Verizon again is some of the most known challenges in the history of business.

Starting point is 01:24:33 Like, good luck. Yep. Yep. And also Verizon's been doing it for a pretty long time now. Although, again, Verizon, the noise around Verizon being like, screw it, you can buy us. It could happen. That might be the solution. It certainly could happen.

Starting point is 01:24:47 All right. I have one more lightning round item for you. And it's about Snapchat, which I like to remind everyone on this show from time to time is a terrible business and a massively culturally relevant piece of software. They announced on their earnings call that there are 971 million monthly active users of Snapchat, which is just a gigantic number. And especially among young people, it is utterly ubiquitous. Snapchat still exists.

Starting point is 01:25:12 Specs are going to be terrible. Snapchat still exists. But Snapchat also did a thing this week that I thought was very interesting and very telling. Snapchat famously like one or two turns ahead of everybody else when it comes to product. Evan Spiegel, the CEO still, I think still on his LinkedIn, calls himself the head of product at Meta, which is a very funny joke that I have always appreciated. It's very good. But Snapchat is now going to attempt to not recommend what it calls wholly AI generated videos in its feeds anymore. And I'm just going to read you a quote from the press release.

Starting point is 01:25:43 It says, as low quality repetitive AI generated content becomes increasingly common across the internet, we want Spotlight to remain a place where people can discover authentic creativity from real people because we believe there's enduring value and rewarding value and rewarding original perspectives, personal storytelling, and the moments that people choose to create and share themselves. You know who agrees with that is everybody, except the companies that are desperately trying to lower the cost of content even further so that it can just all be AI, right? Like, all the stuff we're talking about with Disney Plus and TikTok and YouTube, like, follow that down far enough and it's just feeds of AI content everywhere you look. And there is an obvious cultural revolt against that in progress. And I think it is really

Starting point is 01:26:24 interesting to see Snapchat, which tends to be first to see this stuff and do something about it, make that turn and say, actually, we are going to go out of our way to be the place for humans. Have you seen this rumor that MySpace is coming back, by the way? Yeah. And that its move is going to be, like, as a haven against the algorithmic feeds. Sure. Somebody's going to do it, and it's going to work. And it's really interesting to see Snapchat do it.

Starting point is 01:26:48 Snapchat, I think, has the least to lose because it is so far behind these other platforms. like there's a reason we didn't talk about Spotlight when we talked about Instagram and TikTok, you know what I mean? But also Snapchat is culturally on the right side of history over and over and over again, and I think is going to be again on this one. And it was just fascinating to see them take this biggest swing against it.

Starting point is 01:27:10 All right, and the right side of history over and over again when they just did those glasses is like maybe one step too far. I didn't say every time. I've seen the glasses, man. I saw do dears. I don't know about culturally the right side of history. I agree with you on this.

Starting point is 01:27:26 Not to overdo it, but boy, have we run a lot of stories about how to label and detect AI content. Just Weatherbed, this poor woman on our team, we, like, made her start doing it, made her start covering this like two or three years ago. She knows more about it than anyone. And, like, her point of view is, yeah, none of it really works. And if you want it to work, everyone has to participate. And there are just massive negative incentives to participating. Yes. meta just announced its own weird AI detection labeling system,

Starting point is 01:27:55 and she's like, why? And they don't have a good answer. Like, legitimately, their answer is like, yeah, we don't know why you did this. It's not great. Yeah. But if you want to detect AI generated video, which is quite convincing now, you probably need a metadata solution or a watermarking solution. At this point, it should probably be just Google Synth ID.

Starting point is 01:28:14 People are already trying to figure out how to crack it. The cameras have to support it. The editors have to support it. What do you got to do? You will definitely be able to filter out the little fat guy who, like the, you know, the little blobby guy? You know what I'm talking about? I do. He's adorable.

Starting point is 01:28:30 That's like America's next great IP is that little guy. No one knows what I'm talking about, but I promise you I'll send the clip to our editors and we'll put it in the video feed. Yes, please. But you know what I'm talking about. David knows. I do. And I think it's worth pointing out that the phrase wholly AI generated is doing a lot of work here. You can do an awful lot of AI before something qualifies as being wholly AI generated.

Starting point is 01:28:54 But this is still one meaningful step against this tide of AI. I mean, Google has announced something similar. Like, they all know. They all know it. They all know. But I think Snapchat's saying we're going to do it for real and try to filter it out. They will quickly get to this is a humans only feed. And we're going to have big conversations about AI effects in the back of creator videos or Snapchat's own filters and whether or not they're AI.

Starting point is 01:29:19 But I will say the pressure on the labeling solutions to achieve anything like success is only going up by the day because you cannot make these kinds of promises without metadata or watermarks or labels or something that everyone agrees on works. Yes. And this industry has known it for ages. They have not shipped anything that anyone can agree on or which maybe most importantly of all, which works at scale. Yeah. And there's certainly no chance of Snapchat being powerful enough to make the industry make that turn. The way they would do it is so you can't actually put anything in this feed without said watermark. And that's nuclear.

Starting point is 01:29:55 Yeah. But it might be how they do it. Yeah. It would be interesting. All right, you get one more. What do you want? What's yours? I mean, we promised anger about BMW putting Spider-Man ads in the cars.

Starting point is 01:30:04 And I think we have to deliver. Yeah. You should say the other one is we were going to talk about Open AI responding to Apple's trade secrets lawsuit. Basically, their response was we didn't do it. We'll talk more about that some other time. There's a lot left to come with that story. We should do a whole episode of that. There's a line in that in their reply where it's like, so-and-so who left Apple for opening.

Starting point is 01:30:22 I kept the off-boarding documents just to make sure the other employees he was recruiting followed the proper procedure. So good. It's choice. Instead of breaking the rules, what if he did an even better job of following the rules? He followed the rules so good. You don't even know. He kept a bunch of iPhone prototypes to make sure he didn't build any of them. All right, but BMW, what's going on?

Starting point is 01:30:44 BMW. So there's a BMW integration in Spider-Man Brandy Day. And I haven't seen this movie yet, but I've been told that it's like a very, like, noisy integration. Like, look at my BMW, Spirder. Again, I haven't seen this movie. No, Sputney, not the BMW. That's like more or less the vibe have gotten from people describing it. Kill me, not the BMW.

Starting point is 01:31:06 I've been trying to avoid spoilers, so I haven't really got into it. The actor who plays Ned showed up at the Samsung event. doing Spider-Man stuff. There's just Spider-Man add integrations everywhere. Yeah. Okay. So BMW has this big integration.

Starting point is 01:31:23 And what they have done is they've put an ad inside the infotainment display of many BMWs where if you swipe down, you can see in the app icon there's like a Spider-Man logo that has appeared and you push it

Starting point is 01:31:34 and it plays a Spider-Man informatic. Everyone hates this. There's like a handful of people in our comments on social media that's like, I didn't mind this. If this is how advertising works in the car, it's fine. And then all of their replies are like,

Starting point is 01:31:47 don't do this, it's a slippery slope. Yeah, I do not want advertising in the car. It's a slippery slope. Do not put ads in my car in any way, shape or form. This has such Apple put U2 on out, YouTube album on your phone energy to it, where it's like, you can see why somebody was like, people will love this.

Starting point is 01:32:03 And no one in the room raised their hand and went, no, they won't. Super won't. Not everyone loves YouTube. BMW gave a quote to the Autopian, which is a great side, by the way. These festive animations will be able to be able in more than 70 countries worldwide for vehicles equipped

Starting point is 01:32:16 with the relevant optional equipment and BMW operating system 7-8.59 or BMW operating system X as well as a production date after 7, 2020, which is like, that's a lot of cars or not any cars at all. Like, it's hard to know. Yeah.

Starting point is 01:32:35 Also, I've changed my mind. This is amazing because they just framed we put an ad in your car as an exclusive product. launch. Yes. These festive animations will be available in more than 70 countries worldwide is such a funny way to say, we made an ad.

Starting point is 01:32:52 There's an ad of your car. Our commercials will be available in more than 70 countries worldwide. People truly hate this. I can't tell you how much people actually hate it. It's funny. And on the other side, like Rivian does Easter eggs for Halloween. You can dress up a Rivian as Nightrider. Tesla obviously has Easter eggs.

Starting point is 01:33:10 I think these companies have forgotten that if they want to earn the capital to do this stuff, they have to actually do lots and lots of OTA updates to their cars and like make it fun and like, here's some stuff. And then you can be like, by the way, you're used to getting new stuff in the car all the time. Marvel has made like a Spider-Man and animate. Do you want this? Like it's in the list of all the other fun stuff. And instead they're like, we've done mostly nothing.

Starting point is 01:33:36 The car is still very buggy. And here's an ad. And I don't think they've earned their way into putting these ads in the cars. I'm very curious if a Tesla or Rivian did something like this, so people would freak out as much because at least they have the reputation for delivering new features to their cars all the time. Yeah, and there's a way to do it that this always makes me think of

Starting point is 01:33:54 when they had like the celebrity voices in ways. Yeah. And some of them were very clearly sponsorships of some kind or another. Like if you could like turn your car in the nav into Spider-Man or like one of his cool spidey vehicles. But like you said, that's an option on a list. That's awesome. A thing that just suddenly appears that is like, Spider-Man is here is not that.

Starting point is 01:34:18 There's just no version of that that makes sense. And I don't think that works anywhere. But I think some of these more software-driven cars like you're talking about are just more attuned to how do we build this into the experience. The future of the auto industry, everyone agrees, is soon the cars will drive themselves and you will drive around in a shopping mall of screens. Yes. And we will put advertising all around. to you in the screens because what else you're going to look at? And no one ever in those meetings is ever like, their phones?

Starting point is 01:34:49 No. Seems like the obvious problem with this whole plan. Also, it's worth mentioning, everyone hates this. Everyone hates this. The subtitle of too many of our podcasts needs to be, don't forget everyone hates this. Some people in our comments are like, this was fine. My kids loved it. I will say they put themselves out there.

Starting point is 01:35:08 And I appreciate your point of view. They did get yelled at a lot by other people. But anytime anyone sees a screen with an internet connection, the next thing they see is dollar signs from advertising. And BMW is no exception. Nope. Not great. That said,

Starting point is 01:35:26 BMW, if you'd like to send me any number of vehicles to see the Spider-Man animation for I saw that available. What if instead it just played the TikTok that was the full Spider-Man movie? Just the leak. Would that be better? I did watch a little bit. more of the Spider-Rind movie than I anticipated.

Starting point is 01:35:44 No, Nealai. Just because of the leaks on TikTok. You're going to prison. All right, we should get out of here before you incriminate yourself any further. Eli, who's on Decoder this week? We had a good one this week. It's the CEO of Bose, Lila Snyder. Boz is like they bought Macintosh Labs.

Starting point is 01:36:00 The high-end audio company. They're licensing their tech. We talked a lot about, hey, the future of wearables might be all these, like, AI things. What happens to a headphone company? She got into it. She's great. I really enjoy talking to her. And then next week,

Starting point is 01:36:11 on Thursday, Hayden's going to come back and we're going to go deep into her reporting on the Google AI stuff. Oh, yeah, she's in the mix. She's in the mix. Yeah, that's good stuff. I am excited about all that. Bose, by the way, shipped some new headphones

Starting point is 01:36:23 this week that I'm surprisingly excited about. I love the quiet comfort headphones. Yeah. Like, I have a real weakness for good noise-canceling headphones, which means that I know I'm going to buy these and I'm going to feel very bad about it. Do you alternate between Bose and Sony? I alternate between Bose and Sony.

Starting point is 01:36:37 Yeah. I basically, like, I try to buy sort of every other. generation, if that makes sense. So I'll, like, wear a Sony, a pair of Sony's for two years, and then I'll get a pair of bozes for two years. I also am, like, weirdly hard on headphones for reasons I've never totally understood. I like, I beat up headphones. I've always said you have a big stupid head. I mean, so you and everybody in our YouTube comments says that I have a big stupid head. I also, like, you're supposed to like lovingly put them in a case and then, like,

Starting point is 01:37:05 put them in your bag. And I just like, I don't know why my headphones break all the time. I say as I like huck them at full speed. into my backpack. That's fair. All right. Also, go listen to and watch version history. We're in the middle of making the next season. It's going to be very exciting. That's coming out in a few weeks, subscribe so you get all of it. That's it. We're getting out of here. As a reminder, the best thing you can do to support all of this to make sure that Nilai can keep putting screens behind him, to make sure that I can afford to put something here behind me in this new setup that I have. I move the camera. It's very exciting. It's very exciting. Subscribe with the verge. Theverge.com

Starting point is 01:37:40 slash subscribe. It gets you all of our podcasts add free, including this one and all the other ones we just talked about. It gets you all of our exclusive newsletters. It gets you all of our coverage of everything. Keep an eye on what Hayden's up to with this Google stuff. I think she's going to have a couple more fun stories coming. The goings-on at the top of the AI race feel very important and very in flux right now. Lots to talk about there. And if you have thoughts, questions, feedback, feelings about any of this. If you want to yell at me about my headphone usage, I'd love to hear all of it. Email us, vergecast to the verge.com. call the hotline 866 verse 11.

Starting point is 01:38:13 We love hearing from you about anything and everything. Once I'm gone, there's going to be a lot more hotline on the show. There's going to be like, we're getting back into the regular drumbeat of doing a hotline episode every Wednesday. We're going to drag N Eli into some of those. Lots of good hotline stuff coming up. Get your stuff in ASAP. Vergecast is Verge production and part of the Vox Media Podcast Network. The show is produced by Josh Kajas, Eric Gomez, Brandon Kiefer, and Travis Larchuk.

Starting point is 01:38:36 We will see you next week. I will hopefully see you next week. I'm at some point just going to disappear into paternity. This might be the last one. I'm sorry that I said you had a big stupid head of most episodes. The last time I'll ever see you. I feel really bad about it now. My plan is to be here through the end of next week,

Starting point is 01:38:51 but I may just suddenly disappear into Babyland. And if I do, I'll see you all in a couple of months. Until then, Nilai. Rock in all. Apple News Plus has everything you're into, all in one place. Over 500 publications covering the topics that matter most to you. thousands of recipes from celebrated food publications around the world. Local news from all 50 states.

Starting point is 01:39:18 Sports coverage from across the globe. Audio stories you can take anywhere. And daily puzzles exclusive to Apple News bus. All of it curated just for you. News subscribers try it free for one month at news. com. Terms apply. Support for this show comes from Harvey A.I.

Starting point is 01:39:41 The future of law is agentic. not just tools that assist, but AI agents that navigate complex matters. That's why Harvey created agents that can do the work from end to end. They build a plan, pull from secure data sources, run subagents in parallel, and draft the work product ready for your review. So you delegate the work and on the judgment. Trusted by more than 60% of the AMLA 100 and leading Fortune 500 legal teams, Harvey is the AI operating system designed specifically for legal work.

Starting point is 01:40:12 Harvey AI tailored for law. Learn more at harvey.aI. Support for this show comes from CoreWeave. Everywhere you look, AI is expanding what we thought was possible, and at the center of it all is CoreWeave. Medical research and diagnosis, education, complex visual effects for movies,

Starting point is 01:40:34 science and technology breakthroughs. CoreWeave powers AI pioneers around the world with purpose-built tech, building what's never been built before. CoreWeave is the essential cloud for AI. Ready for anything? Ready for AI. To learn more about how CoreWeave powers the world's best AI,

Starting point is 01:40:52 go to coreweave.com slash ready for anything.

The Vergecast - What's behind the Google AI shakeup

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.