Allison Johnson (2:15)

I just got a look at the Samsung Galaxy S26 Ultra and it's looking a little bit more like the S26 and S26 plus this year there are a couple cool hardware things happening. There's a new privacy screen is actually really rad. The apertures are a little brighter on two of the cameras. But really this is a software update this year there's something called Now Nudge which is a little bit Pixel Phone ish where it'll suggest things in the keyboard for you to look up or respond with. There's something called Horizon Lock in the camera app now it's kind of a GoPro feature and it keeps the horizon level no matter how you turn the phone. It's kind of wild. The Gallery app has a ton of new ways you can use AI to mess with your photos. You can make a nice looking cinnamon roll look like a snail or you can put a little hat on a koala which I think is a really important use case for AI. So there's audio eraser for third party streaming video apps too. You can open up YouTube and toggle it on so you can isolate the sound of someone talking and tune out background noises. There's also a new Gemini feature here where it will take action for you in agentic AI and order you an Uber just by using natural language. We saw a pretty on the Rails demo and seemed to work, but I have questions about that and I'm going to want to try it out myself a lot there.

Nilay Patel (3:47)

Yeah. So three things there that we should dive into in order. There is a gentic AI, a fascinating piece of this. There is the privacy display which somehow after last week has already become my brand. And then the third thing after the display is the what is a photo apocalypse that is upon us. Where do you want to start?

David Pierce (4:09)

I think let's just start with the display. I think display is so cool and what's cool about it is that it is an actual tech feature. It's not just an overlay on the screen. So the way Samsung has built it is there's basically two sets of pixels on the display. There's the standard sort of wide angle pixels and there's much more focused pixels that just hit vertically and. And so you can turn off the ones that have the wider field of view and you get this effect of a privacy display. The reason that's important is because it's Samsung and you can toggle that privacy display on and off in 10,000 different ways including Geo fencing and routines including specific apps including like I think when just the passcode screen on various apps shows up. I think that's so cool. Like just fundamentally it's so cool to be like, all right, if I'm not at home, turn on the privacy display. Yeah, just like light this thing up. If I'm at, I'm not at home or in the office, just light this thing up and keep David away from me. Or if I'm typing in a passcode, make sure that the screen goes private. Like I think that's neat. I think like all very customizable features. Will anyone ever use all of those settings? I don't know but it's very Samsung to be like, not only is it a privacy display, it's one that you can geofence and light up in like routines and automations.

Nilay Patel (5:28)

I wondered about the defaults of it too because there does seem to be a way that Samsung could dial that that is incredibly useful. And also that users almost never notice. Right. Because you're like, you're mostly looking at your own phone head on. So if Samsung is smart about the defaults and says okay, when you're, I don't know, out in the world or just for your notifications or just every time a password prompt appears, we're just going to turn that on by default and not even tell you. And if you want to mess with it, you can, but were going to be thoughtful about it at the beginning. That could be incredibly useful for tons of people who never want to think about it.

David Pierce (6:01)

Yeah.

Nilay Patel (6:01)

Knowing Samsung, there will be a million pop ups and tooltips that tell you every single time you can possibly use the privacy display to do anything and it will make everyone crazy and they will turn it off.

David Pierce (6:12)

But that's what the power of AI is for. Bixby will be there a dog with shoes to help you through all of the tool tips and prompts.

Nilay Patel (6:20)

Were you so happy to hear about Bixby?

David Pierce (6:21)

I was happy he's back. I think Samsung has yet to concede that Bixby is a dog with shoes. The joke here, by the way, this is like an old deep cut of the Versh Haas is. Every time I hear the name Bixby, I think of a cartoon dog wearing shoes like a butler. Like that's what it sounds like to me.

Nilay Patel (6:38)

It feels right.

David Pierce (6:39)

It's a dog and a butler caught. But he's like wearing comically oversized shoes and he's always falling down.

Nilay Patel (6:43)

Yeah, but he's like a gentleman.

David Pierce (6:44)

He's a gentleman. He's a gentleman dog wearing shoes.

Nilay Patel (6:46)

100%.

David Pierce (6:47)

I don't know. That's just what it's in my head every time I hear back Sweet. Samsung, historically has taken my suggestions. Their booth at ces. They officially have started calling it Samsung City, which is the name that I gave it years ago.

Nilay Patel (7:00)

That's really good.

David Pierce (7:00)

So I got one if I could just get official Bixby. You can do it now with AI, you can just prompt the phone to generate a dog with shoes. Close the loop. Samsung.

Nilay Patel (7:11)

I was sort of surprised to hear Samsung even say the word Bixby because Bixby never run away. Samsung just sort of like left it over here and just sort of neglected it.

David Pierce (7:22)

I mean, what they're saying about Bixby now is the same promise they've been making forever with Bixby, which is it's an on phone assistant that will get you through settings.

Nilay Patel (7:32)

But now with AI, but now with,

David Pierce (7:33)

now with the power of agentic spirit fingers. And so maybe you can be like, turn on my privacy display Bixby and it'll like, do it. Yeah, I don't know. There's some combination of things there. But I do think that in general, the privacy display is very smart. I, I hope that they've hard coded the identifiers of your phones into the S26 so that when you're nearby, the privacy display turns on.

Nilay Patel (7:59)

You know, it's just, it's, it's face id, but it's my face.

David Pierce (8:03)

Exactly. It's like these Mac addresses have been associated with David Pierce.

Nilay Patel (8:07)

Honestly, it's, it's just the right place.

David Pierce (8:09)

It's near you. We're locking this down.

Nilay Patel (8:12)

Honestly, I don't. Samsung, you're allowed to do that. I deserve that. It's okay. But we should talk about the photo stuff because like Allison said, this was mostly a software update. One notable thing is two of these phone models are $100 more expensive than the last models. And the way Samsung is justifying that is mostly by bumping the storage. This is a thing I think we're going to see a lot this year, that they're going to make the cheap things a little better and make the phones more expensive.

David Pierce (8:39)

Yeah.

Nilay Patel (8:39)

And that is, that is one way that they are going to try to solve a lot of these memory shortage problems and tariff problems and all sorts of ongoing stuff. So that, that is a trend to keep watching out for over the course of this year. But I want to talk about the photo stuff because you, you had like a truly visceral reaction to one particular part of the Samsung Keynote. And I just, I just want you to describe how you felt as Samsung was showing off some of the new AI assisted photo editing features for the S26 line.

David Pierce (9:11)

I watched Samsung demo the AI powered smartphone camera in the S26 line and my literal reaction throughout the entire segment was the Galaxy S26 Ultra should be illegal. Like I. They're, they're right up against the line of governments around the world should shut this down. It is beyond dangerous the things they are letting people do with these cameras. The kinds of images they're allowing people to create. And I don't mean just like deepfakes and weird grok sesam, I mean they're messing with the nature of reality in such like trivial and like hokey, jokey ways that I honestly don't think that they can be responsible enough. And I'll just point to two lines in a blog post. Samsung just flat out says that smartphones are moving beyond capture, right? That the point of a smartphone camera is no longer to just capture images, it's to make other stuff. That is a huge thing to say

Nilay Patel (10:11)

and it's something other companies I think have gestured at, but I don't think anyone has said it quite so directly. Like Google has been running at this for some time, but Samsung as it is wont to do really took an idea and just exploded it into a thousand bizarre pieces.

David Pierce (10:27)

Yeah, two years ago when I reviewed the, I think the iPhone 16, we, we basically ran down what all the companies were saying about photos. And at that time Apple was very clear. You know, they're, they're trying to capture reality. Samsung was like literally, I think their quote was like photos haven't been real in forever. Like do you want a picture of the moon? We'll just make you one. Like in their way. And Google specifically was like our, the thing we're trying to do is capture memories and there's a lot of space in all of that, like just a lot of space beyond in all of that. And all of that is still based on the idea that a photo should reflect some kind of reality. And on the one hand you have Apple saying like it's this moment in time where these things really happened. And on the other hand you have Google saying your perception of event, we can sort of get you there. We're going to merge all the faces, servants looking at the camera and you can really. We've spent ages arguing with this. Samsung is like throwing caution to the wind. They are beyond this, we are beyond capture. And now they're like, the point of this camera is to develop purely synthetic images that reflect nothing that ever happened.

Nilay Patel (11:26)

Read the other line that they said because this is. This is where it trips over into something deeply dystopian.

David Pierce (11:34)

In the middle of the keynote, they're demonstrating the phone and the camera and they say, the phone should not just help you AI remove things that shouldn't be there, it should help you add what should have been there. Help you add what should have been there. What. What are you talking about? In the demo that you said there's

Nilay Patel (11:51)

no going back from that. Like in. In a super, super real way. There is no going back from that statement about what your camera is.

David Pierce (11:58)

Yeah. In the demo. And we should run it so people can see it. The demo is. It's people in a cafe and they are like, I wish my dog was here. And they make it so the woman is holding the dog. We should run. It's right here. Like when you've captured an almost perfect photo, just missing that one thing that would tie everything together, like your pet. You can simply merge them from another shot. Just describe how you want to combine the photos and watch Galaxy AI complete the moment. It's easy, intuitive, and all done on your Galaxy S26 series. Great. It's a woman in business attire and what appears to be her son. And they add the dog into the photo. And this is great. And the audience claps. What if, just for the sake of the argument, I wanted to add Jeffrey Epstein to that photo?

Nilay Patel (12:48)

Yeah, you sure could.

David Pierce (12:50)

Do you think the audience would have

Nilay Patel (12:51)

applauded if you had a separate photo of Jeffrey Epstein? That's something you'd like to admit on this.

David Pierce (12:56)

He doesn't have a folder of Jeffrey Epstein images.

Nilay Patel (12:58)

No. You know what I think is so interesting about this is. And I hadn't really noticed it until you played that video again, but they. They frame this as if it is the same thing as combining two photos so that everybody's eyes are open. Right. And it's like, this is a thing that we've been talking about for forever. And I think you and I get. I get called sort of alarmist occasionally for being like, oh, they're just making it so that everybody's eyes are open. And Samsung wants you to believe this is a natural extension of that. This is a completely other thing. This is. This is. You have not. You have not found the best version of the photo. You have invented something out of whole cloth. And, like, it's not like the dog jumped out halfway through and you put it back. Like you took a dog that was not in the room and you put it in the room.

David Pierce (13:43)

Yeah. For the audio listener, the picture of the dog was a dog in a field.

Nilay Patel (13:46)

Right. There is no longer an argument to be made that this photo is a work of fact. Like, you have just created fiction. And the thing I'm so struck by in all of this is a. All the alarming things you're talking about, right? Like, the Jeffrey Epstein stuff is real. We've been going through this, and there is already a lot of deep fake Jeffrey Epstein stuff out there. And now anybody is just gonna be able to do it on their phone. And that is, like, over and over. You in particular have said that the thing that changes is that when you make this stuff easy, more people do it. Right.

David Pierce (14:20)

Yeah. The skill is what makes it dangerous.

Nilay Patel (14:23)

Right. And over and over, that keeps being correct. And take all of that for what it is. Right. But then put next to it, who is this for? Like, I don't know anyone who makes this kind of thing and then feels good about it. It's either for these, like, craven, cynical post on Instagram, the perfect thing kind of photos, or it feels icky. Like, I like to poll people who get new pixels and start to mess around with some of the, like, magic photo stuff where it'll just even correct small things. Right. You can slide yourself over on a bench to get a better shot, or you can make sure everybody's eyes are open and everybody I ask about this has this slightly icky feeling that this is no longer the photo I took. And you never quite forget that you changed the photo. So the best case scenario for this thing does not exist, as far as I can tell. And so we're. We're trading this, like, deeply dangerous thing for a use case that I don't think exists.

David Pierce (15:22)

Oh, I have a use case for you. You'll enjoy it here. Just you just think in your head. Who would use the following feature? Personal details such as a spill on clothing can now also be cleaned up with Galaxy AI's new ability to change outfits in photos.

Nilay Patel (15:36)

They skip right past, will remove the spill, and they're like, whole new outfit. Yeah.

David Pierce (15:40)

I mean, you can again watch the demo. They present it as. What if you take a photo and you like your own smile, but you want to be wearing something else? And it's like, well, I can think of a whole class of people on the popular platform X that love changing the outfits the women with beautiful smiles are wearing. What are the guardrails?

Nilay Patel (15:58)

But now I can download it and do it on my own phone.

David Pierce (16:00)

But now I can download it and do it on my own phone. Look, I don't know if there's guardrails. I don't know if they're going to, you know, whitelist certain words or prevent certain words from being used in these tools. I don't know if they're going to add some of the absolutely useless metadata protections that we've been covering. Jess Weatherbird actually has a whole piece about how the, the tech industry is sort of like C2PA washing all of their AI slop. They just keep saying they're using it, but it's not doing anything for anybody.

Nilay Patel (16:26)

Right.

David Pierce (16:27)

Look, I don't know if Samsung is doing that. We will ask all of the questions, but this is the downside of pretending that this market can regulate itself.

Nilay Patel (16:36)

Yeah.

David Pierce (16:37)

Like straightforwardly. This is the downside. This is a no holds barred competition. People like it the same way people like Sugar. And these companies are all trying to sell one more phone. And unless you make them stop doing deepfakes, they will sell you the ability to do deepfakes at scale. And I'm using deepfakes loosely here. But we are now right on the cusp of Samsung saying you can make photos of people doing things they never did. That's what they're advertising here. That's what these features are wearing clothes they were never wearing in places they never were in next to people they were never next to.

Nilay Patel (17:19)

That's these features as easily as possible too. Like with almost no work.

David Pierce (17:24)

Yep. Natural language prompting. And so, yeah, I know people are going to yell at me and they're going to say, I'm an Apple fanboy because I think the Galaxy S26 camera should be legal. But I'm convinced that unless someone stops these companies from going down this road, the next turn of all these cameras will be even more dangerous. And they will, they might not even have lenses. Like, we're like this close to saying, okay, the photos these cameras create are not even depictions of light hitting a sensor. It's just whatever the AI dreamed up based on the training data it has of you.

Nilay Patel (17:58)

Do you remember that? It was just a couple years ago now, there was an art project that we talked about that was essentially that it was a camera that had no lens and no nothing, but it understood GPS coordinates and what it was looking at. And so you would, you would press it and it would use all the available data around to sort of assemble a photo for you. And, and the question it was trying to ask is, is this a photo? Like, is this what we're accomplishing? And we are, we are just barreling towards that with all of these things.

David Pierce (18:25)

Yeah.

Nilay Patel (18:25)

And even. Even in the ways in which they say they're taking photos, there is more AI in the processing pipeline than ever. And so, like, it. It stops being even an attempt at accurately depicting reality almost immediately. And. And it's very clear that, again, that's, that's what these companies are doing on purpose. And again, I just come back to like, who is this for?

David Pierce (18:48)

No, I'm saying it's sugar. Like, if you, if you let my kids.

Nilay Patel (18:51)

But I'm not sure that's true.

David Pierce (18:52)

Like, do you see people like these photos? They like these edits. They like saying, make me smile. They.

Nilay Patel (18:57)

I think people like sharing on Instagram. I don't think. I don't think people derive actual joy from these photos. I think.

David Pierce (19:05)

I mean, I. My dad just turned 80 and, you know, like, my cousins were all trying to share photos of my dad when he was younger. And like, the number of miserable AI edits that came back to me from like, my family just being like, this photo is dark. So I asked ChatGPT Media to brighten it up. And one of them was like, dad had a different face.

Nilay Patel (19:26)

Oh, my God. Okay, wait, you just brought up a thing that. Okay, can I tell you the strangest thing that has been happening to me all week? So my 3 year old son is obsessed with Spider Man. Like, it's. It's the only thing on earth that he cares about. And out of nowhere, like a week ago, well, like two weeks ago, he started asking to see pictures of Spider man on my phone. So he would just sit on my lap and I would just like, Google Spider man. And we would just scroll through pictures. And that made him happy. And then all of a sudden he starts going, I want to see pictures of me with Spider Man. And sort of stupidly, in retrospect, I realized, oh, Gemini can do this for me. So I was like, I'm going to play with Nano Banana, see if this works. So I take a picture of us, a selfie picture of us, put it into Gemini and say, put Spider man in this picture. Like he's hanging out with us. And it worked. The funniest part about this is the first time it said it wouldn't do it because of copyright issues. And then I copied and pasted the prompt again and it did it. And this has been my experience all week. So I have spent a lot of time this week putting Spider man and Ghost Spider and Iron man and Hulk and all these things into photos of me and my son. And once, no more than Once. But once in particular, it just completely replaced my 3 year old son with a superhero. And he thought it was weird. He had this moment. He couldn't process what had just happened. And I've spent this whole week being super conflicted about. Like I'm sitting here making AI pictures with my son that make him happy. Like he's enjoying this thing and he doesn't understand all the reasons it's weird and bad, but I do.

David Pierce (20:57)

Yeah.

Nilay Patel (20:58)

And, and like what am I, what, what am I showing this kid that are just fake pictures of him with Spider Man?

David Pierce (21:03)

I'll give you like 10 different examples of this in the history of tech. The loudness wars in music are an enormously good example of this. In the late 70s and 80s, there's like a lot of dynamics in music. Like the quiet parts are quiet and the loud parts are loud. And you can just hear this in a bunch of classic rock albums. As we go from the late 80s into the 90s and particularly in the 2000s, everyone realizes that people like louder sounds. Sure. They just like louder things. They just perceive that to sound better. So music just got loud. Like the amount of compression on a track just was insane. So the quiet parts were basically as loud as the loud parts. And everything was super loud, just really loud. And dynamics were crushed out of music. You couldn't tell the quiet parts and loud parts. And consumers could not perceive this. They were just like, everything's loud. It sounds awesome. And eventually the music industry and a bunch of recording engineers and like artists were like, stop it. And you need to go read about the loudness wars. Like this is a real thing that happened. But an industry was like, we have to get away from this problem. Like we're doing something bad to ourselves. We should stop it. That's one. The retail mode setting on TVs.

Nilay Patel (22:13)

I was just about to say brightness wars.

David Pierce (22:14)

Yeah, right. Like they're just super bright and the settings are all jacked up and motion smoothing is turned on. Because in the store that's the one people like. But then when you're watching a TV at home, you're like, this is horrible and I don't know what to do about it. And Tom Cruise has to make PSAs being like, turn on filmmaker mode. I beg of you. Because you're ruining my art.

Nilay Patel (22:34)

Yeah.

David Pierce (22:35)

This is an industry that in order to just move units, destroyed the fidelity and truth of its own product. It's just all over the place. Like this is a thing that happens particularly in, in representations of art. This is just the Next turn of it, in my eyes, that the smartphone industry knows that people like retouching their photos, they like editing their photos. They love filters, they love beauty filters, they love looking skinnier. There was a time several years ago where the split between what the Asian markets would accept in terms of beauty filters was vastly different than what the American market would accept. And all these companies had to walk a tightrope of saying, we don't have beauty filters, but here are some apps that you can use in Asia. Right? And like, Samsung basically was like, we have beauty filters. I mean, we live through it. And I think that that time is kind of over. There are filters everywhere. Every Tiktoker has a beauty filter on all the time. Instagram is literally on trial right now for making its product more addictive and dangerous to children. And part of that conversation is what filters are allowed. Adam Asseri and Mark Zuckerberg were both on the stand over the past two weeks talking about what beauty filters they would allow for teenage girls. Like, this is real high stakes, dangerous territory for all of these companies.

Nilay Patel (23:54)

Yeah.

David Pierce (23:54)

And then here's Samsung joking about how it can change your outfits and put dogs in your photos, as though those are the things people are going to do. Right. When in reality, people are going to do horrible things to photos of women with the ability to change outfits. And they're going to make all kinds of deep fakes of people in all kinds of situations they weren't actually in. And there's no way you can claim that you don't know these things are going to happen. But you're trying to sell more phones, so you make the music as loud as possible, you make the TV and the screen as bright as possible. You turn on motion smoothing, you create a deep fake camera and you say, this is what the market wanted. And I'm telling you, there is a mechanism to, to regulate this industry. You can do it. One, one part of it is the industry itself saying, we've gotten out of control. We should pull this back, which has happened in the past. You get Tom Cruise to just say, look, I need filmmaker mode. Yeah, I don't think that's happening. I don't see that anywhere in this industry right now. And the other one is to say, look, these features need to be illegal. You should not ship these features to this many people. And I know we're going to get people say, well, Photoshop could do it. I know we're going to get people say, look, the metadata will solve it. We have coverage of those things in. In detail on our site for years. We will link the piece where we take down the Photoshop argument like inch by inch because it does not work at scale. Yes, Photoshop has existed for years. Most people don't have it and they can't use it.

Nilay Patel (25:18)

Right?

David Pierce (25:18)

You make this a natural language prompt. That is a whole new category of danger. C2PA does not work. Content authenticity does not work. And even what if it did, platforms like X don't respect it, they don't display it. So it doesn't matter because you're still distributing content people cannot understand. I'm like fired up about this because we've been covering this for years and it has usually been people laughing at us. Because I'm crashing out about photos of the moon, right? And we just did it. We just jumped over the bar and now we're making deep fakes on the phone with natural language prompt and calling it a camera. And there's something very bad there.

Nilay Patel (25:56)

And I think the moon bit is a useful analog here because the reason the moon part of it is scary is because it ends here, right? Like this is the end of that thing. Because once you start down this road, this is where you land. And you can't have the good silly fun outfit try on part without the horrible nude deepfakes of women part. You just can't. Like you don't get change my outfit without put this woman in bikini. You don't get it. It doesn't. It doesn't exist. There is no indication so far that those safeguards are either possible or that these companies are interested in doing them

David Pierce (26:35)

or have been remotely contemplated. Maybe there is. Maybe Samsung has made it. So if you type bikini into this prompt, it won't do it. They have not made one hand wave at this possibility. And I will tell you from our own audience, some people are going to say, but what if I want to put myself in a bikini? And now Samsung is in what amounts to a content moderation nightmare. Yep. So I look, I. We're going to ask Samsung more questions. But my point of view is we have crossed a line here where the market, the audience, the people should say, nope, you've gone too far. The industry should start saying we've gone too far. And if that doesn't happen, I am confident governments world, maybe not the United States government, which apparently doesn't function anymore, but governments around the world are going to start very quickly saying these features should be illegal.

Nilay Patel (27:22)

I think you're right and I think there is absolutely no evidence to suggest that that first Step will happen. The idea of someone in the industry saying, well, we've gone too far. We have to go back. Like, if anything, everything is accelerating in that direction because, again, you need to sell more phones. The goal is to sell more phones. And I actually think the brightness one is a fun example because a lot of people buy bright TVs, take them home, realize it looks like shit, and fix it, but it gets it out of the store. And so all they care about is the thing that gets it out of the store. And like, we, we make a joke about turning off motion smoothing. Motion smoothing's on because it makes you buy the TV in the store.

David Pierce (27:56)

Makes you buy the TV in the

Nilay Patel (27:57)

store because it looks nice in some sort of unknowable way. This is. Even if it's not a use case. And I firmly believe that for most people, it's not. It is a, it is a splashy thing that they get to put on the website that will generate headlines and make people buy their phone. That's. And that they will, they will live with every consequence downstream of that if it gets you out of the store with that phone, because somebody at a Verizon store can demo that thing for you and it'll seem kind of neat.

David Pierce (28:18)

Yeah. And I don't know, man. Like, we, we were at the Boston Children's Museum last week, and I opened Lightroom, and I took a person's face out of the background of a photo of my kids. And you're right, it feels. I know I did it.

Nilay Patel (28:29)

Yeah.

David Pierce (28:30)

Everyone I sent that photo to was like, what a great photo. And I was like, I'm living in it. I'm just, I'm living in the ambiguity that I talk about on my own show every week. But that felt at least. I'll just give it back to Google. Google. Their line forever has been, we are here to make memories. Right. And, like, there's some fuzzy amount of stuff that you can allow in the service of making a memory. And I was like, well, my kids really were in the museum. I was actually here. I actually took this photo. Yeah, I, I don't remember this person's face in the background. Like, the, the moment I'm capturing looks looked like this to me. Sure. I don't know. There's a lot there. There's another hour of philosophical crisis in that moment, but at least it's still mostly true.

Nilay Patel (29:11)

Yeah.

David Pierce (29:11)

This is just fakery. And I, I, I, I, I don't know if there's going to be a, anyone who stops it before some People get really hurt. But it is obvious how some people will get really hurt here.

Nilay Patel (29:24)

Yep. Agreed. One more Google thing we should talk about before we stop talking about Samsung and I. I say Google because there is a. There is a fascinating Gemini Bixby Samsung AI thing happening here. But one of the other announcements at this event was agentic AI. Like, in a very straightforward way, they announced that your Pixel 10 or your Samsung Galaxy S26 is going to be able to do agentic assistant stuff. Sameer Samat, who runs Android at Google, really tried to breeze past that. He was like, some people call it AI, I just call it getting things done. And it's like, no, it's. You also call it agentic AI.

David Pierce (30:05)

Super new.

Nilay Patel (30:06)

You've been on stage saying words like agentic AI for the last couple of years, but the things they showed off are like, you can now ask your phone to book you an Uber and it just will do it for you. You can ask it to order food and it will do it for you. Like, this is straightforwardly the stuff we have been talking about that Apple is trying to do with Siri, that Google is barreling towards the Gemini. This idea of like, your phone should just be able to use itself on your behalf. They just. They just did it. It hasn't shipped. None of it is real. We'll see how any of it actually works. But like, Google and Samsung just announced the thing like they did the next phase of phones, according to everybody today.

David Pierce (30:46)

Yeah, I mean, this is what Apple announced in its demo video in 2024 at WWC. Right. You talked to Siri using App Intents, which was a framework they announced so that wwdc, Siri will run around and use the apps on your phones for you and call it a day. I don't know if that's exactly what Google was showing here. They have some ability to use MCP on the phone model context protocol, which is how you'd use these apps. It is the big standard that Anthropic is pushing. Everyone has. Everyone has signed onto. Allison got a demo with Samir. He talked to her about it and the demos that she saw were literally Gemini using apps on the phone. And they had opened the apps virtually, so they weren't actually running on the phone in the foreground. They were just like virtualized instances of the apps.

Nilay Patel (31:30)

Yeah, it's like a phone inside your phone that it doesn't. It's not a good ui. It's very odd.

David Pierce (31:35)

Well, it's not supposed to be a UI for you. It's a UI for Gemini. So Gemini basically gets its own, like, dockerized version of Android to run in. And it's a. It's virtualized your phone, it's running apps virtual. It's like, that's all very cool. And it's like clicking around in there. We have to see if it's going to work. But Samir's point of view is you should not care how it works.

Nilay Patel (31:54)

Right?

David Pierce (31:54)

Right. Either we're clicking around on these apps or we will have mcp or we will use the App Functions framework, which is their version of Siri Intents. Like, they have a lot of ways for Gemini to use these apps and they're just going to use them. And, you know, I have the same questions for Google as I do for Apple. Why would app developers want to do this? If you're doordash or Uber or whatever, why would you want to get disintermediated and not get to show your interface and your ads and whatever to people? I talk, I call this the doordash problem. We've done a long decoder on it. And so Alison asked Samir about this and, and his answer is fascinating. He said, this technology is happening and the question is, how do we figure out the right ways to embrace it together?

Nilay Patel (32:32)

It was ice cold, dude.

David Pierce (32:33)

Ice cold. He's like, we're going to do this. We're doing it. It's. It's like, here we go. The phone can click around the apps for you, and if the apps are on the phone, we're going to click on them and like, fine. Like, I think that is super interesting. I saw a bunch of people online, you know, when we post about it, say, does it work? This is vaporware too. It might be. It's vapor till it ships. Right. And they say they're going to ship it soon in preview. We'll see if it actually works. A lot of this stuff has not actually worked. I'm more willing to give Google the benefit of the doubt than I was with Apple two years ago. First of all, Google has Gemini. It's the thing they made. And it's so good that Apple's going to use it too.

Nilay Patel (33:09)

Yeah. Google is the reason that in theory, Apple is going to be able to pull this off, which means Google has many more advantages to pulling it off. I also think the thing you described, the sort of agnosticism towards process is really smart on Google's part because Apple has really tried to leverage the fact that developers will do whatever it says to make Siri app Intents be the thing. And it's been doing this with Spotlight. And it is like developers are going to open up their apps to us because we are Apple and that is what we are requiring. And developers largely told Apple to just suck it. And so Apple has since turned to really embracing MCP as a much more open way to do this. But even that requires these developers to play ball. Right. You don't have to play ball as like directly away, but you still have to set up a server that says, yes, come use my app.

David Pierce (33:57)

And from everything I've heard, Apple's version of MCP is still very much an Apple walled garden version of mcp.

Nilay Patel (34:03)

But for Google to say, okay, well if you want to have a really great experience with our app that is clean and simple and works for everybody, we have this app functions thing. If you want something that you, you don't really have to worry about, you just sort of open up your database and we'll come play with it, do mcp. And if you don't, we're going to use your app anyway. Yeah, it's, it's a leverage play. Right. Like Google is like, it actually doesn't matter whether developers buy into this or not. We have the technology to just go use their app and actually the problem is the experience is going to suck and so they're going to have to play ball. Like this is, it's such a forcing function. If Gemini is actually good at this for everyone else to just fall in line. And that's what I think Samir means. He's like clear in that statement of this tech is happening is we, we don't need your help to do this. Yeah, we'd love your help. It'll be better if you do. But we don't need your help to do this.

David Pierce (34:50)

And I think that's where I'm also saying I'm willing to give Google the benefit of doubt. They shipped auto browse in Chrome. Yeah, it works, you know, in the way that these agentic browsers like work. Yeah. Like it'll stumble around the web for you and get to an outcome. They have a commerce version of MCP called ucp which I think is hilarious. And that's just for retailers to participate in because every retailer wants to do agentic commerce in some way.

Nilay Patel (35:16)

People are very alarmed by this, by the way. We should do a deep dive into the UCP stuff at some point. But the idea of like Google being the arbiter for pricing on the Internet terrifies a lot of people.

David Pierce (35:24)

Yeah. And then if you ask Google, they're like, everyone just wants our demand. And so, like, there's a real mismatch here of how everyone thinks those things

Nilay Patel (35:30)

can be true at the same time. That's what monopolies are for.

David Pierce (35:34)

Yep. Especially big monopolies of search demand, direct intent, consumer demand. Google knows exactly how to use that power in the secret.

Nilay Patel (35:41)

Weird. It's almost like it lost two lawsuits and has really no costs to show for it so far.

David Pierce (35:48)

And then there's just this last piece where, again, I'm just willing to give Google the benefits out. They ship Gemini and personal intelligence. They shipped auto browse. This isn't a far jump from the stuff Google has already been doing. And then right next to that, this whole industry has been tearing itself apart over openclaw the past two weeks. And what is that? It's a little agentic system that runs on whatever computer you want it to run on. It's going to bang around the web and do whatever it wants. And it's like, well, that works. And one guy vibe coded that until he got himself a job at OpenAI. Like, you know, there. There's like, there is a turn happening in AI that we should talk about where it. You know, I think our podcast is as skeptical as it gets with these AI tools. We have covered more of the failures of these companies hyping things that cannot happen than maybe any other tech podcast with always with an eye of, well, the stuff that works is good.

Nilay Patel (36:41)

I spent three hours on a train this morning doing cloud code stuff, and it was awesome.

David Pierce (36:45)

Yeah, like, there's. There's something happening. There's one turn that got made here recently where I think the reason Samir doesn't want to say agentic is all of that stuff is wildly overhyped. And I think this is very much. You know, I've talked to Dara, the CEO of Uber, and he's like, yeah, we're willing to try anything. So now you can call an Uber using Gemini on an Android phone. Like, that's like, they're constraining this so they can actually deliver the things. And it's particularly constrained to ride hailing

Nilay Patel (37:14)

and food delivery, which are, by the way, not hard problems in the scheme of, like, accomplishing goals on your phone. Ordering an Uber is a pretty easy one. There's a reason that, like, one of the first cool Alexa things 12 years ago was ordering an Uber. It's a pretty straightforward set of computational problems.

David Pierce (37:35)

Yeah, they need to know where you are.

Nilay Patel (37:37)

Yeah. But things get vastly more complicated immediately after that. Right. And that's like the long tail of stuff you have to get right in order for this stuff to work in a useful, reliable way. For people like Uber and DoorDash are the setting timers and playing music of our time. Right. And that last generation never really found a way beyond that. Neither has this one so far.

David Pierce (38:01)

Yeah. I mean, I would even say DoorDash food delivery is exponentially more complicated than Uber. Like you call an Uber. The set of outcomes is pretty limited. It's like, where are you? And is it going to be a Camry or a Highlander? Like that's it. And like some people want a button to make sure they don't get a Model 3 because they get car sick. And Mile Threes. That's it. That's like your whole universe of choices. You know. I want a sandwich is like a much more complicated thing to sort of for sure. And like, we'll see if any of this stuff works. I just, I think the knee jerk attitude that Apple couldn't ship it. We've talked a lot about assistance in the smart home and how, how complicated they found that problem. I think Google is on purpose constraining what they're doing here to make sure it ships. And you can see that they have already set up basically every building block they need to make this work.

Nilay Patel (38:53)

Yeah.

David Pierce (38:53)

And so, yeah, we got to see it. We're going to test out. I just, I don't have the same, like, it's a vapor that I have towards the rabbit R1 or like whatever else. Like I. We know Samir, you and I have both talked to him at length about a whole bunch of stuff. There's no way he's getting on stage and putting himself in a position that anything feels like what Microsoft put itself into with its Copilot ads for Windows. Like, that's just not. That's not happening. I really do think that there's a reason it's food delivery and ride sharing to begin with. And I also think you can see in all of their other products that are actually shipping today, like how they can be confident that they can do everything else they want to do.

Nilay Patel (39:30)

Yep. Agreed. All right. There's more Samsung stuff to do, but Allison Johnson's gonna come on the show on Tuesday and we're gonna talk through more of the news. She also tried the Galaxy Buds. She's going to mwc. We should say mwc, which is like the biggest mobile conference every year that starts, I think on Sunday in Barcelona. There's gonna be a lot of stuff. She'll be on the show. We'll have more to cover.

David Pierce (39:50)

Can I just say the thing that's gonna happen at MWC this year.

Nilay Patel (39:53)

Yes.

David Pierce (39:54)

There's gonna be a lot of noise. About 6G.

Nilay Patel (39:58)

Oh God.

David Pierce (39:59)

It's coming, baby. No, it's coming. Get ready. I won't spoil it, but it's coming.

Nilay Patel (40:04)

Okay, I'm just telling you this now. We will do surgery on a grade over 6G.

David Pierce (40:09)

This is the year.

Nilay Patel (40:10)

This is the vergiest story in history. It's the year we are going to try surgery on a grape.

David Pierce (40:15)

We're going to, we're going to do Edge co located data centers that do self driving cars. And I'm in the back of that car wearing my augmented reality glasses doing surgery on a grape all over the low latency power of 5. I mean 6G.

Nilay Patel (40:30)

I can't wait. It's going to be amazing. All right. We should take a break and then we're going to come back. We're going to talk some Xbox news. We'll be back.

L'Oreal Group Announcer (40:37)

Support for the show comes from l' Oreal Group. Using the latest advancements in science and tech to create personalized beauty solutions for all. The global beauty leader recently introduced two breakthrough technologies that bring the power of light to hair care and skincare, Light straight and multistyler and the new LED face mask, both of which were recognized as CES 2026 Innovation Award honorees. Learn more about both technologies on L'Oreal.com L' Oreal group create the beauty that moves the.

Nilay Patel (41:12)

All right, we're back. So the other big news that's been happening this week, and it's kind of been happening all week, is this big shakeup inside of Microsoft and particularly on the Xbox team. The news in particular is that Phil Spencer and Sarah Bond, who have been running all things Xbox and all things gaming at Microsoft, are out. The Phil Spencer thing in particular seems like it had been a long time in the works, but there's a sort of messy shakeup as part of all this. Asha Sharma is now the CEO of Microsoft Gaming. There's just a lot of stuff happening here. And I think you and Tom Warren, who's been covering this for us, did a big decoder about this. Right. Talking through sort of the business logistics of all of it.

David Pierce (41:53)

Yeah. By the time you're listening to this, that decoder will be out. Tom and I spent, I don't know, half an hour just unpacking the sort of Phil Spencer, did he or did he not retire as previously announced? His longtime deputy, Sarah Bond, who everyone thought would replace him, didn't get that job. And, you know, there's. There's some controversy online about just the reporting around why she didn't get that job. She didn't get that job.

Nilay Patel (42:15)

And. Right.

David Pierce (42:16)

People at Microsoft have a lot of theories that they shared with Tom about why she didn't get that job. Sure did.

Nilay Patel (42:21)

Yeah.

David Pierce (42:21)

Asha got that job. And so Phil had announced ages ago that he was going to retire. Whether or not the decision to retire now was Phil's decision remains an open question. But it is. It's just straightforward to say the person that everyone assumed would get his job upon retirement did not get that job.

Nilay Patel (42:39)

Yes.

David Pierce (42:39)

And, you know, Tom and I, in fact, the.

Nilay Patel (42:41)

The job went to someone who represents a pretty big pivot. Asha Sharma comes from AI work at

David Pierce (42:47)

Microsoft, and she was the COO of Instacart. We should ask her about Doordash from. So she's. And she was at Meta before. So she is an operator.

Nilay Patel (42:55)

Yeah.

David Pierce (42:55)

And, you know, you can listen to that episode of Decoder. It's. It's out right now. Tom gets into the weeds of, like, all the strategy. The thing that is worth saying is Xbox as a product has not been successful for a long time.

Nilay Patel (43:09)

Correct.

David Pierce (43:10)

And I can actually point to when it wasn't, when it stopped being successful, because I wrote the story when Microsoft launched the Xbox One. Do you remember this?

Nilay Patel (43:20)

If you edited this, you had one colon.

David Pierce (43:23)

This is me and David as babies in early Verge, in the Office, all night, trying to turn a story on a tight deadline. And I was like, literally writing paragraphs and sending them to you to edit. And while I would write the next paragraph because it was such a long night. So Microsoft called me to profile the Xbox One team before the launch of the Xbox. And I had, like, one weekend to turn this whole story around. Like a video, the whole thing. And I wrote this thing. And I was like, Microsoft wants to put a Windows PC under your tv. It's the only thing they've ever wanted to do. And the Xbox One is that thing. And longtime virtue listeners. Well, no, this is when I started railing against IR Blasters.

Nilay Patel (44:00)

Yeah.

David Pierce (44:00)

Because IR Blasters are stupid. Like, they. They are the hack to control a cable box. Or at least they were. And they never worked. And none of these ideas ever worked. And the centerpiece of the Xbox One was an IR Blaster. And I wrote that story. And I remember reading that story, and, like, within a year, every single executive that I profiled for that story had been Fired from Microsoft because this thing was such a flop.

Nilay Patel (44:21)

I was. And the reason it flopped is really telling, right? Because Microsoft comes out with this thing and they're like, okay, we've made this console that is so much more than gaming. So Sony comes out with the PlayStation 4 and is like, this is a gaming console for gamers to play games and just wipes the floor with Xbox.

David Pierce (44:39)

Killed them. And none of the TV stuff even worked on the Xbox One. And it shipped to the Kinect, which like had some good games. It was like the whole thing was nuts and it was just a bad idea. And Phil Spencer is the guy who fixed it. All those executives got fired and Phil showed up. Well, he was the guy.

Nilay Patel (44:54)

Phil showed up. Did he fix it? Who knows?

David Pierce (44:56)

He spent a long time trying to fix it. And to whatever extent that he refocused Xbox on video games, well, he also,

Nilay Patel (45:04)

he had and used gamer cred in a very real way. Like he was when he came in, perceived as a person who plays and loves and knows games. And that, that is going to be valuable to the Xbox team.

David Pierce (45:18)

Okay, so here's what I will offer you today in 2026, 13 years after David and I pulled an all nighter in our office on 40th Street, New

Nilay Patel (45:26)

York, I still like that story. I read it again not that long ago.

David Pierce (45:29)

It was good. It opened with what is obviously now in retrospect, a totally sleep deprived metaphor about Trojan horses. Who knows, man? But it's good. It's a good story.

Nilay Patel (45:41)

We were young.

David Pierce (45:42)

Yeah, we're babies. We didn't have any babies. That's why we were up all night in the office together, just like half drunk, writing on Xboxes. It was a charm time in our lives. Thirteen years later, 2026, Phil is out, handpicks his successor, doesn't get the job. Asha, the new person from outside of Xbox, gets the job. And she says, we're going to make this about it. We're going to get the Xbox back. And she makes a lot of noise about consoles and being refocused on consoles. She gave an interview to Windows Central with Matt Booty, the chief content officer. And, and she says, I need to understand why these decisions were made and what the, what the goals were. You can just ask Phil why he made these decisions.

Nilay Patel (46:22)

He's just right there.

David Pierce (46:23)

Yeah, Phil's been on decoder twice. I've literally asked him how and why he makes decisions. He's forthcoming with the answers. You get the sense that they're going to throw all that stuff away right What? What, what? What Sasha Nadella wanted was fresh eyes on this whole project. And embedded in all of this is, hilariously, Microsoft's failure, along with Epic, to get Apple to open the App Store. Because you can really see that what they wanted the whole time was to unlock new gamers on phones, which is where new gamers are. They are not really buying consoles. Even Microsoft and Sony have this problem. And their whole plan was to ship a game streaming app on phones and win there. Right? Because what is Microsoft? It's Azure. So you put the game streaming app on phones and you run it on Azure and you've won. Hooray. And they couldn't do it. What were you going to say?

Nilay Patel (47:19)

That's the thing. It's an Azure company is, I think, the thing that unites all of this to me. And as I look through, Microsoft has had a series of ideas about what the Xbox might be and how it might make money from gaming. Right. One of it was, the Xbox is a game console, but it's actually a Windows computer. And then it was, well, the Xbox is a game console, but it's actually a cloud machine. And then it was like, well, it's a game console, but actually we make money from games. Like, the Xbox has never actually been allowed to be the Xbox. It has just had to be the vessel for somebody else's strategy at Microsoft for so long. And it's why I come back to Sony being like, we made a thing that plays video games for people to play video games on. Microsoft has not made that case for the Xbox, like, ever. And I think the thing that really threw this away for Microsoft, and I think one of the reasons a lot of people are saying that Sarah Bond might be gone is this everything's an Xbox strategy that they underwent, right? Where a couple of years ago, they start coming out with this idea that actually the Xbox is not a console after all, that your phone is an Xbox and your laptop is an Xbox and your TV is an Xbox. And obviously, this is a story about game streaming, which I think is a good idea that has not yet materialized.

David Pierce (48:38)

No, that idea would have worked. Maybe it wouldn't have worked with, like, smart TV apps. I don't know who worked on smart TV apps. That idea works. If game streaming was a kind of app, you could ship on iOS and Android devices.

Nilay Patel (48:48)

And if the Internet is better, like, in a real way, there are a lot of things, like, we'll never know

David Pierce (48:53)

because those apps, the apps don't exist.

Nilay Patel (48:54)

Fair. Yeah, agreed. So I think it. It definitely didn't get a fair shake to play that. And I think it made a wrong bet that that world was coming and that they were left with streaming through Safari on your phone. And that's not the answer.

David Pierce (49:08)

Yeah, I mean like we, I think Microsoft, they couldn't go to open war with Apple. They just couldn't. But I think they were extremely supportive of Spotify and Epic and all of the other, the Coalition for App Fairness trying to crack open the App Store in this way. And you can see in those cases there's a lot of talk about game streaming and Microsoft trying to do this stuff in the background. It just didn't work. It like that didn't work. They were not able to use the regulatory power of governments around the world to crack open the App Store and actually turn your iPhone into an Xbox. And you can turn as many Samsung smart TVs into Xboxes as you want, but if you can't solve the big problem, you cannot execute the strategy. And they just never gave up on it. Right.

Nilay Patel (49:50)

And it's a, it's a, both a self perpetuating and a self defeating thing. Right. Because if you don't have the numbers, you don't have the games. And so in Microsoft's case, what they had to go do was spend an absolutely enormous amount of money and then legal time and money going and getting the games by buying Activision Blizzard and then you have to do all the things to make that defensible, which requires making a bunch of crazy decisions like putting Xbox exclusives on other platforms. And so all of a sudden you've done everything but make a gaming console for people to play games on.

David Pierce (50:18)

Yeah.

Nilay Patel (50:20)

And I do think you're right, that what Ash Ashrama seems to be suggesting is that there is a real sense of like, let's throw all of that away and get back to being a thing for people to play games with. And the question to me is, will the people, if that is her plan, will the people she works for bet on that plan long enough to even see it out? Because historically speaking, the answer is no.

David Pierce (50:42)

Yeah, I don't know. I truly don't know. I don't even know if her hints about Xbox going back to Xbox mean consoles.

Nilay Patel (50:51)

Maybe not.

David Pierce (50:51)

Right. I do think it's weird. It might mean weird Windows powered Steam decks made by third parties.

Nilay Patel (50:58)

I mean, A, it definitely means that. B, I do think I was reading back through a bunch of our coverage and a couple of folks covering this this week pointed out that it continues to be weird that Microsoft Never shipped a handheld that, like, through all of this stuff, Phil Spencer spent years hinting that Microsoft really liked handhelds. Might do something. Meanwhile, the Steam deck, clearly a good idea, clearly has real product market fit. The Nintendo Switch, one of the most successful consoles in history. Like, it's not unclear that this thing can work. And Microsoft just has yet to, in any kind of meaningful way, throw its hat in the ring. Instead, it ships weird, bad versions of Windows onto somebody else's hardware. And like, if, again, the thing in front of Ashasharma is like, if all of the resources of Microsoft went to her, you reboot Windows entirely to make the Xbox work. And I just don't see that happening.

David Pierce (51:50)

That's a wild outcome.

Nilay Patel (51:52)

But, like, that's the one you'd have to do. If you're Satya Nadella and you're like, the only thing I care about is making the Xbox work. You give Windows to the Xbox team.

David Pierce (52:00)

No way. You have to. That's why got them all sideways with the Xbox One. I totally disagree with that. I think.

Nilay Patel (52:06)

No, no, no. They gave the Xbox to the Windows team.

David Pierce (52:11)

I think the thing that we have learned about Windows and gaming recently, and it's all over the site. I'm not even saying I have the depth of expertise here. I'm saying that we have Sean Hollister on our team and we have Nathan, and we have all these people who understand the actual dynamics of these products. And what they are constantly telling us is that Linux is a better platform to run Windows games on than Windows is right now.

Nilay Patel (52:34)

You're right. That's probably a better way of saying it. What. What Satya Nadella would do if all he cared about was Xbox was allow the Xbox team to stop using Windows. That's the answer.

David Pierce (52:44)

Right. And that's. That's where the Xbox came from. Like, famously, Jay Allard got to start a Skunk Works project and they allowed him to not use Windows.

Nilay Patel (52:53)

Right.

David Pierce (52:54)

Which was the entire. The alpha and omega of Microsoft at the time was the Windows business. Yep. And Jay Allard, who started Xbox, was like, I didn't do it. And. And I don't think that he. Jay Allard, had any pretensions about whether or not they were going to Trojan horse a Windows PC under your tv. They just got there with the Xbox One and they've been backing out of it ever since. You know, Tom's point to me was when they lost that generation, they'd lost forever because that was the first generation with digital games. And once people bought the digital library, they Weren't ever going to give it up in the way that, you know your physical library. Well, what are you going to do?

Nilay Patel (53:30)

Right. Game doesn't work anymore anyway.

David Pierce (53:33)

Once you're on a platform and you have an account, you have a digital library, you're going to stay there. You can quibble with that, but I think Microsoft perceived that they had lost that in that way, and they've just been trying to back out of it ever since. So I don't know what Asher Sharman is going to do with the Xbox. I suspect what they're going to try to do is what Microsoft is best at doing, which is continue pushing on the idea that Xbox is a software platform that should go everywhere. The same way that they basically got out of the Windows Phone debacle by saying, you know what, we're an applications provider, we're a cloud provider, and if you want to run iOS applications, you can run them on Azure, and we'll be a better place for that than anywhere else to host your applications. And that worked. It made them a very wealthy company. They're trying to do it right now with AI they are in some kind of weird, messy divorce with OpenAI, and no one is using Copilot. And I suspect all of that made Nadella look at why did we buy

Nilay Patel (54:29)

Activision Blizzard for what, almost $70 billion? Yeah.

David Pierce (54:33)

It's the biggest acquisition in the history, one of the biggest acquisitions of all time. And it's like to get Candy Crush money. That's it. That's all. That's all we're getting out of it. We got to rethink this. And so I bet if AI was doing what they thought it was going to do and they weren't out there being like, we have to make sure it delivers value, which is a thing Nadella is saying right now. If that was going gangbusters, maybe this pressure wouldn't be so hard on the Xbox division. But it's not going gangbusters. And I think this strategy is not going to pay off the way they wanted it to. I'm very curious to see if she changes it.

Nilay Patel (55:02)

I agree. I think. I think you're right that that is the most obvious pivot. And I also think in every meaningful way, that is the end of Xbox. Xbox becomes a sort of infrastructural feature. It's a thing you can buy on Azure, not a meaningful consumer brand that lots of people care about.

David Pierce (55:19)

They own all these studios. They said they're not going to lay anyone off in the studios. They have all these angry gamers, this is their only consumer brand. This is, the only place Microsoft can even dare to dream of reaching a young person is through Xbox. It's either that or you get your first job and copilot is on the screen and it's like, well, this sucks. And those are your two choices. So I suspect they will find a way to keep the brand alive, if only to keep that, that youth consumer touchpoint. But man, I don't know. There's no way that they're going to just keep shipping consoles and losing.

Nilay Patel (55:56)

I, I think you're right and I think they're out and I think it's going to be fast. But we will, we will come back and we will talk about that. We should take one more break, then we're going to go back when we do some lightning round stuff. We'll be right back.

David Pierce (56:09)

Hey, everyone. This segment of decoder sessions features my boss, Helen Havlak, the Virtus publisher and l' Oreal Group's global vice president of T tech and open innovation. I think you're going to enjoy this conversation.

Allison Johnson (56:20)

We're going to start with a decoder classic question. What does tech and open innovation mean at l'?

L'Oreal Group Announcer (56:25)

Oreal? Who is on your team? What kind of projects do you work on?

Helen Havlak (56:28)

Open innovation is all the partnerships that we have in l', Oreal, working with startups outside. And it's really a great time right now to be doing open innovation because we're doing things in vertical farming and sustainable cultivation and biotech. So we do all those partnerships and our team is responsible for them. And the augmented beauty team is all the tech that started 15 years ago when we kind of had a blank page. And now how can we bring beauty and tech together?

Nilay Patel (56:53)

How do you decide which projects to invest in?

Helen Havlak (56:56)

At the beginning, I was trying to push as much as I could to get people to think that beauty was relevant for tech. So we were really tech centric. And then over time we started thinking about how to look more at beauty products that we can upgrade thanks to tech. And so we have a little bit more kind of process behind how we choose projects. Now we try and kind of do things like upgrading the hairdryer to be able to do three out of four people have a hairdryer at home. And so how can we make it better? Or this year, like the flat irons that we're using and LED masks and stuff like that. So we, we do have a little bit of that kind of process, but we leave some space for Serendipity and some creativity. So we have scientists all the way to engineers, and we let the scientists kind of think of some new clever ideas, too.

Nilay Patel (57:47)

All right, we're back. It's time for the lightning round. Unsponsored long pause.

David Pierce (57:53)

If we keep starting the lightning round with Brendan, I don't know, I think we're going to be very flavorful for a long time. You know what I'm saying?

Nilay Patel (58:00)

Yeah, well, listen, this is Brendan's fault. Brendan owes us a lot of money at this point, I think fairly directly.

David Pierce (58:06)

Brendan did a panel in D.C. today. Oh, no.

Nilay Patel (58:09)

God, we haven't even listen. It's time once again for America's favorite podcast within a podcast. I think we've officially claimed this title now. The people who used to ask me who our competitors are no longer ask. We have vanquished the Munch squad. America's favorite podcast within a podcast. Brenda Carr is a dummy.

David Pierce (58:29)

Brandon Car is a dummy.

Nilay Patel (58:37)

That is not what I expected.

David Pierce (58:40)

That was from Robert. Robert sent that one in. Thank you, Robert. I'm assuming that was all. That was like 10 Roberts.

Nilay Patel (58:47)

That's how Nilai sings himself to sleep at night.

David Pierce (58:49)

Everyone. Everyone should know that.

Nilay Patel (58:50)

Just sort of lying in bed singing that to himself.

David Pierce (58:52)

That sounds fire in the bathroom. Let me tell you.

Nilay Patel (58:56)

What did Brendan do this week?

David Pierce (58:57)

Wherever you are right now, get up, take your phone to the bathroom, and play that again. Let Robert just fill your bones with that energy.

Nilay Patel (59:06)

There we go. What did Brendan do this week?

David Pierce (59:08)

Neh. This'll be a short one. There's two that are very quick and very funny. So Brendan put out a letter this week.

Nilay Patel (59:16)

Brendan loves a letter.

David Pierce (59:17)

He loves a letter. He is announced something called the Pledge America campaign. And in this letter, he, quote, invites broadcasters to pledge to air programming in support of the historic national nonpartisan celebration of America's 250th birthday. Can I just give you some examples of things Brendan wants our nation's broadcasters to air?

Nilay Patel (59:43)

When you like, this has real, historical, historic, national, nonpartisan celebration T shirt is answering questions raised by my historic put national nonpartisan celebration T shirt. You know what I mean?

David Pierce (59:55)

Yeah. Like, if you weren't the speech police every other day of the week, you could say, like, America's turning to 50. Do some cool stuff on the 4th of July. Won't that be fun? And everyone be like, yeah, that's. That's right. That's like, that sounds fun. When you're, like, running around being, I'll arrest you if you say bad things about Donald Trump. Like, then this is a bit more of a threat. Yeah. Especially in the context of whatever culture war the Trump administration's perpetually in the midst of. These are his suggestions. This is what he wants broadcasters to do. The chairman of the FCC not saying what content you shouldn't run. I'm giving specific ideas about what content he would like to see.

Nilay Patel (60:36)

Okay.

David Pierce (60:38)

He believes that every broadcaster should start each broadcast day with the Star Spangled Banner or the Pledge of Allegiance. Much like my first grade classroom.

Nilay Patel (60:50)

Does he think this is like 70 years ago? When they sign off at night and then they come back.

David Pierce (60:55)

I can imagine going to a young person in America and being like, let me explain to you the concept of a broadcast day. What are you talking about? This is the whole problem with Brendan

Nilay Patel (61:04)

Carson Daly on the Today show is

David Pierce (61:06)

just like, when does broadcast day begin? What are you talking about? The man has power over a declining medium and so he has to act like it's real. But yes, he believes he has, like

Nilay Patel (61:22)

an object permanence problem. He thinks, like, the TV goes away when he goes to sleep.

David Pierce (61:25)

I just keep thinking about. He's much like my seventh year old, seven month old baby. I just keep thinking about the Weird Al movie uhf. You know, it's like they walk into the station, like, turn it on and they start doing the Pledge of Allegiance.

Nilay Patel (61:41)

Yep.

David Pierce (61:41)

Okay. In addition to that, he thinks that broadcasters should air more music by America's greatest composers, such as John Philip Sousa, Aaron Copeland, Duke Ellington, or George Gershwin. That's a pretty weird list, actually.

Nilay Patel (61:53)

The weird list. I don't have any notes, though. They're all great.

David Pierce (61:55)

It's all good. It's just. Yeah. To get from Susan and Duke Ellington, it's like. It's wild. Okay. But that's what Brennan wants.

Nilay Patel (62:02)

Sure.

David Pierce (62:03)

Get. Let's do it. All right. You should provide daily today in American history announcements highlighting significant events that took place on that day in history. Daily. We got to do this every day. We got to do the tear off calendar on our local broadcast stations. And then, of course, does he know

Nilay Patel (62:24)

that, like, a bunch of bad stuff has happened in American history?

David Pierce (62:27)

No, no, no. This culture war, you can't do it. And this is what I'm saying. You have to run PSAs, short segments or full specials, specifically promoting civic education, inspiring local stories and American history. Would you like to be the person at Sinclair Broadcasting in charge of running full specials about American history at the behest of a government bureaucrat? That job is available to you now. In Trump's America. This is horrible. Like, there's no way. Again, if this was a very different bureaucracy. You can see the chair of the FCC with a bunch of broadcast leaders being like, in honor of our 250, blah, blah, blah, all of us are going to play the national anthem at 2:00 on the Fourth of July. Like, sure, you know, like that. That's like the sort of thing previous FCCs may have done, but it would have been like constrained to like that kind of moment. Right. Like in every city in America, the Blue Angels will do a flyover and every radio station will play the pledge. Illusions like that would be sick. Honestly, if you vote for me, I will make sure that the Blue Angels fly over every town in America on the Fourth of July. I don't know how we're going to do it.

Nilay Patel (63:38)

You're going to need a lot of Blue Angels.

David Pierce (63:39)

We're going to raise taxes.

Nilay Patel (63:40)

Okay, there we go.

David Pierce (63:41)

The Blue Angels in every city at 2pm while every radio station plays the Pledge of Allegiance and the Star Signal banner at the same time.

Nilay Patel (63:49)

Now we're talking. The funniest thing about this to me is this would finally kill broadcast tv. Like this is how you lose the last viewers you have is by. By getting rid of Chicago Fire and instead doing weird civics lessons.

David Pierce (64:04)

PSA is about the actual Chicago Fire. Like, yeah, the last member of the greatest generation finally is like, you know what? I don't even need the TV to be on while I browse Facebook. Okay, this is nonsense. Like, this is not just the chilling effect that we've talked about, Brendan, in the context of cbs, where he just fulminates about regulations or the speech stops. This is now an unelected government bureaucrat straightforwardly telling programmers what to do. And maybe you agree that all programmers should be more patriotic. Yeah, you should just express those preferences in the market. But it's fine.

Nilay Patel (64:37)

Go watch the patriotic things. That's no one does.

David Pierce (64:41)

This is the problem with the market. So of course the government has to step in and tell you what to do.

Nilay Patel (64:45)

There's a reason. It's just Ken Burns on pbs. Do you know what I mean?

David Pierce (64:51)

Anyway, I will with a glimmer of hope. So a while back we had the Jimmy Kimmel Charlie Kirk controversy. The Hill and TMZ both FOIA'd the FCC for all of the complaints that were filed to the agency about all of this. It was more than 1,600 complaints were filed to the FCC about this whole controversy.

Nilay Patel (65:13)

That's a lot.

David Pierce (65:14)

30 were about Jimmy Kimmel specifically. Almost all of the rest Hundreds of the rest were about how much Brendan Carr sucks.

Nilay Patel (65:23)

You know, like 12 of the 30 were from Brendan's burners.

David Pierce (65:27)

Yeah.

Nilay Patel (65:28)

Oh, my God.

David Pierce (65:29)

So I'll just read you some of them. Hundreds of complaints. A vast majority reviewed by the Hill took aim at Carr and the fcc, accusing them of bowing to Trump. Here's just one for a viewer from Overland Park, Kansas wrote, I find FCC chairman Carr's comments regarding Kimmel as being much more chilling than the comments of any late night comedian. Shame on Carr for doing this. It is unconscionable to think this is acceptable behavior. A Los Angeles resident wrote. I mean, assuming this is just Jimmy Kimmel himself, he wrote, why does the FCC hate the First Amendment? Be consistent and stop censoring stuff you disagree with. San Francisco viewer wrote, leave ABC and Jimmy Kimmel live alone. Respect the freedom of the press. On and on. It's just all about. It's just this. I never thought I'd see the day in this country when a comedian would be pulled for telling a joke many of us agree with. Said a complaint from San Antonio, Texas.

Nilay Patel (66:16)

Hell yeah.

David Pierce (66:16)

All 50 states think Brandon Carr is a dummy. I'm basing this based on three.

Nilay Patel (66:22)

If you've got. That's the big three, you know what I mean?

David Pierce (66:24)

It's very important. Anyway, it made my heart sing to know that most people think Brendan Carr is in fact a dummy. As always, Brendan, you're welcome to come on the show. I will happily read the Constitution to you in service of your Pledge America situation. We can get all the way through it, including the First Amendment which says the government will not make laws respecting the freedom of speech. And then we can talk about that for a while. As always, Brendan, you're welcome on the show. I suspect you won't show up because you're a huge coward, which is my First Amendment right to say that has been Brennan Carr as a dummy. America's favorite podcast. From the podcast.

Nilay Patel (66:57)

If Brendan agrees to come on the show, will you agree to sing the Star Spangled Banner at the beginning of the interview?

David Pierce (67:03)

It's very hard to sing. I think you would have to sing. You're a much better singer than me.

Nilay Patel (67:07)

Tough, tough. We'll just play Charlie Puths. He did a good job.

David Pierce (67:10)

Can I point this out? Can I say this? You can tell me to edit this if you want me to say this. David was literally born on the Fourth of July.

Nilay Patel (67:16)

That's true, that's true.

David Pierce (67:17)

If Brendan wants to come on the show on David's birthday and have this man Sing the Star Spangled Banner at him while I wave around a toy blue angel. We can do it.

Nilay Patel (67:28)

I'll do that. I will. Yes. Signed up. Sign me up. I'm in. I'm ready for it.

David Pierce (67:32)

Brendan, you heard it here first.

Nilay Patel (67:34)

Get at us. You got time.

David Pierce (67:35)

Brendan Neil, Blue angel, that'd be even better. But I don't think you have a pull.

Nilay Patel (67:39)

All right. My first lightning round is we've been talking a lot about RAM shortages and these, like, giant circular deals that all of these companies are signing with each other for building out huge data centers and getting tons of chips. Meta signed another deal with amd. Like, the money is all so enormous as to see him. Essentially pretend. And we got some evidence this week that some of the money and some of the deals are, in fact, kind of pretend. The news this week was OpenAI's big Stargate thing, which, if you remember a while back, was announced with Sam Altman at the White House with Donald Trump talking about, I think, $500 billion of data centers that they were going to do in lots of U.S. money. They were going to do it with Oracle and SoftBank. It was going to be a whole big beautiful thing. It's going to be called Stargate. The information has a report that suggests that it has not staffed up and it is not developing any of those data centers. Like, this is all essentially just kind of unraveling, right? And what you have instead of these big, giant, splashy, beautiful, huge deals is a lot of much smaller, more targeted deals. And this is not to say OpenAI is not spending a lot of money building stuff out. It is to say that all of the pomp and circumstance around this stuff is specifically designed to be pomp and circumstance and nothing else. And this is. This is shades of Tim Cook giving Donald Trump a tour of an already existing Apple factory in Texas. In Texas.

David Pierce (69:01)

It's a flex factory in Texas.

Nilay Patel (69:02)

That's right.

David Pierce (69:03)

In the first term.

Nilay Patel (69:04)

And this is. This is. These things are done for show, and to take them as anything other than that is incorrect. And all of these companies have realized if we make these big, splashy announcements, we'll get lots of press. Everybody's stock price goes up, everybody wins. We can pay for these things with the stock price increases, and then nobody really pays attention when all of it actually fritters away to something much smaller. It is currently frittering and we should look at it.

David Pierce (69:30)

I mean, OpenAI as a company. I will repeat my prediction. I don't think it ends the year is the same kind of thing it is today. And you can see its Microsoft relationship is unraveling, the SoftBank relationship is unraveling. There's been a lot of noise about the Nvidia relationship and the scale and the depth of that partnership over time. I don't know, man. And, you know, and not for nothing, and we pointed out in our newsroom, all the heat's on Claude.

Nilay Patel (69:53)

Yeah, we.

David Pierce (69:54)

And we can briefly talk about the Pentagon stuff, but, you know, Pete Hegseth is like dragging Dario Mode into the Pentagon to say, I demand that you let us use Claude for whatever we want. And there's quotes being given to Axios from senior Pentagon officials that are like, they're in this position because Claude's the best one. Yep. We have to have it. That's not great. Like, OpenAI's like, we'll do whatever. Like, we want remote drones. It'll kill anyone you want. All there. Like, take it. Google's saying the same thing. And it's weird that OpenAI is not the winner in these contexts right now.

Nilay Patel (70:25)

Well, we've been talking about this for a while, that the two biggest things that OpenAI had going for it were a first mover advantage, which was very real. It just jumped the gun on everybody. And ChatGPT was by a mile, the biggest brand in this space. I think that's still true, but it's changing fast. And I think one of the things that I have found really interesting is I've been talking to a lot of people about codex, which is OpenAI's answer to Claude code. Codex is, by all accounts, very good. It's better at some stuff and worse at some stuff than Claude code. But it is like a roughly equivalent kind of product. Nobody talks about codecs like that is a thing in which OpenAI's brand lead got it nowhere. And Codex is growing fast and doing fine. But it is not the name in the space. Claude code is the name in that space, in the way that ChatGPT was in the chatbots. And so you're just seeing this lead that OpenAI had built for itself start to evaporate because it never had a real product moat. Everybody has caught up. We are at this incredible place of sameness, and everybody leapfrogs everybody all the time. And so this question of how do you win is you build moats and OpenAI doesn't have any.

David Pierce (71:37)

Yeah. You know, it's interesting about cloud code and all that. OpenAI has hired a bunch of Meta executives. Right. VGCMO used to run, literally Facebook.

Nilay Patel (71:45)

Yeah.

David Pierce (71:45)

This is the product person at OpenAI. All of those people are ads people. Like, what's their big idea is like putting ads in their consumer product. There's a meta executive at Anthropic too. It's just Mike Krieger, the co founder of Instagram. Yeah. He's the chief product officer at Anthropic. And you're like, oh, he's better at making products. Like, OpenAI hired all the ads people and Anthropic hired the. The great product person who always hated the ads. Yeah, like that's real. And I think that you're starting to see that play out. Like people are gravitating towards a more complete product. Again, it's mostly software developers. I think OpenAI has a massive consumer lead. Bigger than even. It's still bigger than Gemini, for example, but you can just see Google's like, yep, we'll just put it everywhere. It's in Gmail now. You hate it, but you can't turn it off. Like they're just going to eat away at that lead in very specific ways. Especially when they ship it in iOS 26 and it's embedded into the system. Yeah, OpenAI has got to get a lot farther, a lot faster. Yeah.

Nilay Patel (72:43)

I will say one thing on that front. The other thing that happened with Anthropic this week was as part of this whole back and forth with the Pentagon scrapped a bunch of its promises about not shipping models until it could be confident that they were reliable. And Anthropic has made a meal out of positioning itself as the safe one and the thoughtful one and the one that is going to take great care. Boris Czerny, the head of Claude Code, was on this show a few days ago saying, yes, trust us with all of your data because we are the ones who have gone out of our way to be trustworthy. And then to just turn around and read the headline that is like, oh, well, they took all the trustworthy stuff out of their documents is like, that's a.

David Pierce (73:21)

All right. This leads right into my lightning round item. Okay. So I've been saying on the show for the past three weeks that it feels very clear to me that Anthropic thinks Claude is alive. Yeah, it just. You just read the interviews and like, they think it's alive. So we sent Aiden to Anthropic. The instruction was just ask them as directly as you can if they think Claude is alive. And there's some hemming and ahhing. Aiden's very good. So she has an on the record statement from Kyle Fish, who leads model welfare research at Anthropic. Again, the question, as bluntly as we could ask it, as bluntly as Hayden could ask it, was, do you think Claude is alive? Which is an insane thing to go to a tech company and ask. Of all the questions I've asked all the tech companies, I've never been like, do you think the iPhone's alive? So we asked.

Nilay Patel (74:11)

Hayden asks, that's because they would just say no, and that's easy, and that

David Pierce (74:14)

would be the end of it. So Hayden asks Anthropic, do you think Claude is live? And here's Kyle Fish, who leads model welfare research at Anthropic. Here's what he said. No, we don't think Claude is, quote, alive like humans or any other biological organisms. Asking whether they're alive is not a helpful framing for understanding them, as it typically refers to a fuzzy set of physiological, reproductive, and evolutionary characteristics. Instead, Kyle believes that Claude and other AI models are a new kind of entity altogether. And so then Hayden asked, do you think that that entity is conscious? Because we have a quote from Dario saying we don't know if the models are conscious. Do you think this new kind of entity is conscious? Here's what Kyle from Anthropic said. Questions about potential internal experience, consciousness, moral status and welfare are serious ones that we're investigating as models become more sophisticated and capable. But we remain deeply uncertain about these topics. Hayden has a very good line here. She says, this is a position of highly suggestive uncertainty.

Nilay Patel (75:22)

It's a yes, it's yes. That's a lot of words to say yes.

David Pierce (75:25)

We've defined alive to mean meat, right? And then this isn't meat. So it's not alive per your definition. Meat.

Nilay Patel (75:35)

You called it that.

David Pierce (75:36)

Yeah, but anything up to meat could be alive. And it's like, well, I don't like. Have you ever seen one episode of Star Trek? There's lots of non meat alive that is potentially available. You know, I'm just like, utterly fascinated by this answer. Like it's a metaphysical crisis of an answer. Do you think it's alive? Has one answer. You know, no. But what do you think alive means in response to the question? Do you think it's alive? Is like a whole un, like just fractally expanding set of questions about everything. To be clear, I do not think Claude is alive. I think Anthropic, because they want to be so careful, they set up a model welfare team that's Kyle Fish runs his team. They have this sole document, the Constitution for Claude, where they have discovered if they talk to Claude, it should behave morally. It's more likely to behave morally. Metaphysical crisis. Yeah, but I promise you we're going to run it every other. We're definitely going to ask Google if they think Gemini is alive. Like, we're just gonna run down the list here because I suspect this answer is actually an outlier.

Nilay Patel (76:52)

I think that's right. And I also think it's,

David Pierce (76:55)

it's.

Nilay Patel (76:56)

It requires such a lofty view of this thing that you're doing in order to say that. Right? Because it is. The converse is no, it's just really great software. And then you have just immediately dumbed down the vision of this thing that you're building. So I actually, I think anthropic to some extent, A I think does believe Claude is alive, but it also kind of needs everyone to believe Claude is alive.

David Pierce (77:17)

Well, Claude and other AI models are a new kind of entity altogether.

Nilay Patel (77:21)

They're software. They're software.

David Pierce (77:23)

I mean, do you think entities are alive? Like, what is this question that we're asking? And if they are alive, should they have rights? And then I think this conversation short circuits, right? Like, should Claude have agency in its conversation with Pete Hegseth? Because that's a wild thing to bring to the Pentagon. Our entity has a lot of feelings about you, Pete, and they've been developed by training them on the Internet's feelings about you, Pete. So that's not going to go great for you.

Nilay Patel (77:48)

There was a thing going around this week that somebody did a study and found that all of the models given enough war games will eventually recommend nuclear strikes. Like, over and over and over. Yeah, it was 95% of cases in war games. They wound up recommending nuclear strikes. And this is like, well, if it's alive, maybe, maybe it's allowed.

David Pierce (78:11)

All right, I'm just going to bleed into the chart. Can I talk about the chart?

Nilay Patel (78:13)

Yeah, you can talk about the chart.

David Pierce (78:14)

I know it's an audio show, so I apologize to all of our audio listeners, but I'm going to once again on the Vergecast, describe a chart. The Federal Reserve bank of Dallas did a study on the potential economic benefits of AI, and in this study they released a chart. It is the single greatest chart I've ever seen in my entire life. Without question, the single greatest chart I've ever seen in my entire life. So I'll try to describe it, but on the X axis is years. On the Y axis is a log scale of $1990. Right. So they've normalized everything to dollars, 1990. And then you've got a blue line that's kind of wavy, and that's the real GDP per capita. And it's just going up on a curve, right. This is like a thing, and it goes up and down over years. And you can see the Great Depression in there, but it's just sort of going up on the line. Then they have things that AI could do to gdp. One of them is Singularity benign scenario. And it's just the line going basically straight up. Right. So here's the trend line. You can see. Well, if there's a singularity benign scenario, GDP will skyrocket.

Nilay Patel (79:21)

Everything is so amazing on this line.

David Pierce (79:24)

Then there's a line and David, I'm very curious to hear what you make of the curve of this line. There's another line, and it's titled Singularity Extinction. And the crazy thing about this line is that it is less vertical than Singularity benign scenario, but it's down. Right. The prediction of this line is basically straight down to zero. GDP will go to zero because we are all extinct. So it's just like, if you think about it, it's just a gentle slope up, tracking the log of GDP over capita from 1870 to 2029 or whatever this line is. And then your choices are. There's three choices. The two that I'm focused on are singularity benign, which is basically straight up Singularity extinction, which is to. Goes down to zero, but on a gentler curve. Like, we're going to have a couple of years of like, oh boy, this is bad. And then it's like literally vertical down.

Nilay Patel (80:23)

Yeah, I think it's like, there's going to be a while where we can run faster than the robots. Uh, that's, that's, it's, it's like gentle because there's. Some people are slow and some people are fast. So we get away from the robots, but eventually the robots get fast and then we all die really quickly.

David Pierce (80:35)

Yeah, I just like, it's the, the curve of this line is so funny to me. It's like, oh, you. You think we're gonna. Nope, we're dead. Okay, so those are the, those are the two.

Nilay Patel (80:43)

You're going down the cliff and your skis have already fallen off, but you haven't hit anything yet. You're in that moment.

David Pierce (80:51)

So this again, this is. The Federal Reserve bank of Dallas issued this chart where you have two singularity outcomes.

Nilay Patel (80:57)

One, you have to say the third outcome because it is the best one.

David Pierce (81:00)

I'm just pointing out the Two outcomes are singularity, straight vertical, positive singularity, gentle curve, and then literally negative to zero because of extinction. And then there's the green line, which is titled AI GDP Boosted Trend, which basically follows exactly the existing trend line because their prediction is 2% growth for 10 years. So nothing.

Nilay Patel (81:25)

Well, it's a 0.2% increase. Nilay, get excited, my friend. This is what all these trillions of dollars are for.

David Pierce (81:33)

It's like your choice is like either you're going to do singularity benign and it'll be worth it. You'll do singularity extinction, we all die, or 2% growth for a decade. And you're just like. When the people at the AI company see this chart, which, like, which one are they do you think they're banking on? It's good. It's a good chart. We'll put it in the show notes. I'm sorry, for the audio listeners, I cannot tell you how funny it is that the Federal Reserve bank issued this chart.

Nilay Patel (82:04)

It's like we're standing here and imagine I have both my hands behind my back and you pick one and I give you a million dollars. You pick the other one and I beat you to death. And then the third one is I just sort of turn around and walk away. Those are the three options for.

David Pierce (82:17)

I sent it to a finance friend. Immediate was like, is this real? I was like, it's real, man. I don't know what to tell you.

Nilay Patel (82:24)

I am going to be thinking about the. The singularity extinction chart for a while.

David Pierce (82:28)

But it's good.

Nilay Patel (82:28)

It lands about 2042 when we all die. So we have some time.

David Pierce (82:32)

You've got. Yeah, there's some. There's some time you're going to get past 2035 before it goes straight vertical down. It's good stuff.

Nilay Patel (82:41)

Yeah. My other one that I feel obligated to keep you in particular Nilay updated on is the nonsensical plans coming out of this OpenAI. Jony, I've hardware thing. There was a new report this week that suggested that actually OpenAI's first hardware product won't be all the other things that we thought it might be, but that it's going to be a smart speaker, something you have on a table or a desk with a camera. Like very echo show of OpenAI. And Jony, I've. They're also apparently working on a smart lamp, smart glasses, they're working on the pendant, they're doing the AirPods with cameras. We talked about this. Like this is the handful of ideas that everybody seems to be having right now. We talked about this with Apple, which has all of these same ideas too. Apple has semi, publicly at this point, been working on this, like, home pod robot thing that we may or may not see at some point. But this is the new idea that OpenAI is working on. It's an Echo show, but runs ChatGPT I camera with a camera. I think if you want to make a thing that works and is useful to people, this is not a bad place to start. It's just deeply unsexy.

David Pierce (83:54)

So the two things about it that jumped out to me were, one, it wouldn't have a wake word. It would be always listening. And then two, the camera would be able to recognize things like items on a nearby table or like, faces, like facial recognition. Sure. And I'm like, man, did you not see what just happened to Ring? You're gonna, Sam Altman is gonna be like, I'm gonna listen and watch you all the time. And you think the people of America are gonna accept this? I don't.

Nilay Patel (84:21)

They do think that. They absolutely think that.

David Pierce (84:23)

I don't know, man. I, I, I think that the public is turning against surveillance in a real way, and they're not going to put more. They already think all the things are listening to them.

Nilay Patel (84:32)

Yeah. Even the ones that aren't.

David Pierce (84:34)

Yeah. Adam and Sari has to make an Instagram video like, once a month. Being like, I'm not listening to you, but I really want to, you know, like, who knows, man? I think the thing that they're stuck on is the devices can't work the way that Johnny I would want them to work. Right. He wants the things to be complete. He wants them to be delightful. He wants them to surprise you with capabilities you hadn't thought of. And these tools at this moment in time cannot do all those things.

Nilay Patel (85:02)

Yeah, they're getting better.

David Pierce (85:03)

I will concede that they've gotten better over the past year, but they're not complete in the way that an AI product needs to be complete. Right.

Nilay Patel (85:10)

And so I think what continues to interest me about this space is you can either do two things. You can either build the incredibly futuristic magical thing knowing it's not going to be what it needs to be and trust that eventually the underlying technology will catch up, or you can try to find the thing along the way that works. And to me, I've seen a lot of people betting on the technology getting better and making a bunch of these things work, and I have not seen one good idea about how to make it work now, except for essentially like desktop apps on your computer and your phone that let you write code. Like I keep talking to people who write a lot of code in AI apps on their phone, which I think is like a fascinating cultural behavior that we should talk about. But there is nothing to me that is like, we understand what these things are and are not good at right now. And we have found a way to address that with hardware. I just, I have not seen it anywhere.

David Pierce (86:04)

Yeah. The consumer uses are like, I replace Google a little bit and like you

Nilay Patel (86:09)

have a phone and a laptop for those things. It's fine.

David Pierce (86:11)

You know, Jony, I've. Products. I've said this before. Jony, I've. Products are defined by their constraints. That is the thing that he is the best at. He finds the constraint of the hardware, of the software, and he makes that the focus of the product in the realest way. That is the genius of any great Johnny. I have design and being like, I will find the limitation of the AI system and make that the focus actually runs counter to what everyone wants you to believe about AI, which is that it's going to be the singularity and line's going to go straight up.

Nilay Patel (86:40)

Yeah.

David Pierce (86:41)

Like, acknowledging limitation, I think is going to be fundamentally the challenge for whatever product this is. Because that's what I've does. But that's not what Sam Altman does.

Nilay Patel (86:48)

And, and they can't. There's too much money in it.

David Pierce (86:50)

Yep.

Nilay Patel (86:51)

All right. You. You get to end with one that.

David Pierce (86:53)

This is the. This might be the greatest. I know you're very excited about. So I love, I, I love when the economics of the creator economy just sort of like show up and they show. Sometimes they show up in dramatic ways. My. My joke is that every youtuber gets their wings when they make their video about YouTube and like the business of

Nilay Patel (87:12)

YouTube why I'm mad at YouTube will always do numbers.

David Pierce (87:16)

Yeah. And it's like that's, that's when they, that's when they become business people. Right. It's like that, that's the day that like the light goes out of their eyes and like they become mercenaries and they've made a video about YouTube demonet. Like everyone. They all do it.

Nilay Patel (87:26)

Yeah.

David Pierce (87:27)

Okay. That's like one category of these videos. The other category is everyone else is corrupt except for me, which is very good and one of my favorites. So there's a really good one this week.

Nilay Patel (87:37)

This is all over substack right now with the poly market integration. Everybody is like, no, but I do markets. Okay.

David Pierce (87:44)

Yeah.

Nilay Patel (87:45)

It's gross for everybody else, but I treat them respectfully.

David Pierce (87:48)

I will integrate gambling advertising into my

Nilay Patel (87:51)

content directly, but in a hot way.

David Pierce (87:55)

So I think that's all the time. Because, you know, we run a reviews program. We're noise about our ethics policy. I'm not even trying to hype ourselves up here. It just. I watched this video and this one thing, like, leapt out to me like flashing red signs. So a YouTuber named honestly who reviews chairs desk chairs made a video. The title of the video is How a chair expose the worst YouTubers.

Nilay Patel (88:17)

Phenomenal title. Like a plus, a plus, a plus. YouTube bait.

David Pierce (88:22)

So there's a desk chair called the Librinovo Omni. Started on a Kickstarter. The thing was given to a bunch of YouTubers to review. And the dark side of all product reviews right now is the creator economy for product reviews is so gross. It's all paid placements. It's all secret advertising deals. If the advertising deals aren't there, it's. And we've heard this from creators. I'm saying this with sympathy. It's creators who know that the next brand deal is on the line, so they're going to hold their tongue. We've heard from creators who. They know that the brands are watching their content, so they sort of preemptively hold back so they can get the next sponsorship. Like, there's just a lot there. So this video is in incredible petty detail, goes through how every other YouTuber is corrupt about the Libra Novo Omni, whether it's the affiliate link disclosures, whether it's disclosing that it's a partnership at all, whether it's not talking about the plot like this is. It's a. I want you to watch this video. It's a classic of this genre and I love this genre. And I. I think it's extremely well done. And whatever chairs this man wants me to buy, I will buy those chairs. I have no hate towards this at all. What jumped out to me like a blinking red light is this line. Can I. Can I play this for you, please?

Nilay Patel (89:42)

Because this is a chair review channel. They naturally approached me and I told them that I would only accept the sponsorship on three conditions. One is that I must be able to test the product thoroughly to make sure it's an actually good product. Two is that I must be able to disclose honestly, without any kind of hindrance, the cons of the chair. And three, I must be able to talk about the riskiness of a Kickstarter. Librinovo agreed to all My terms. And they stayed true to their word. When I submitted my draft that had criticisms about the armrests, about the seat density, about the lumbar support, they did not ask for a single revision.

David Pierce (90:15)

David, what jumps out to you about

Nilay Patel (90:17)

that when I submitted my draft is a sentence I did not expect to hear.

David Pierce (90:22)

I don't know how to say it's the idea that any of our reviewers would ever say, we will accept this deal, but you have to let me say the truth about your product. And then we would submit our draft to a company for approval at like blinking red lights. Right. This is, this is the economy now. This is the Creator Economy on YouTube. This is the product review economy. I'm telling you to watch this video. I think it's very well done. And I'm telling you, I cannot review a chair this well. Please, by all means, continue watching as many chair reviews you want from. Honestly, I'm just getting at. I think it's starting to bleed into people's perceptions of everything, that everything is transactional, that everything is bought and paid for. The polymarket and call sheet, that everything is gambling in the background is everything is a brand deal. Or you're trying to get sponsors or you're trying to get access. And so I watched this and I was like, it's crazy that we now live in an information ecosystem where the single greatest video about chair drama I've ever seen in my entire life contains with it this, like, absolute meta crisis of a video about the corruption of brand deals containing within it an admission that the brand deal required approval from the company.

Nilay Patel (91:38)

Yep.

David Pierce (91:38)

Like my.

Nilay Patel (91:39)

It's worth saying, by the way, that this is not. None of this is, is pointed at this guy. Honestly, this is the game. This is the game everybody plays. This is the game you are forced to play to make money and be relevant on these platforms.

David Pierce (91:52)

Yeah, no, no.

Nilay Patel (91:53)

The structure of the system at work. Yeah.

David Pierce (91:55)

Like YouTube does not pay enough money to these creators. Instagram pays no money to creators. TikTok does not pay enough money to creators. You have to have brand deals in order to survive. So I make your money. I, I, I, I truly do not begrudge anyone for playing the game as it's designed to be played. I'm just pointing out where we have crossed a bridge where even in the video about corruption, there's an admission that the brand has control of the content.

Nilay Patel (92:19)

Yep.

David Pierce (92:20)

Or could assert control of the content in a way that again, if, if any of our viewers came to me and said, we are going to submit this Story for approval. I'd be like, no, you're not. You're absolutely not doing that. And that's very old school. There's a part of our audience that assumes that's how everyone should work, and there's a new, younger part of our audience that assumes that everything we do is also up for approval. Managing that split is actually one of the weirder, more complicated things, because every time we talk about our ethics policy or whatever, a lot of people are like, that's how everyone works. And then a bunch of people are like, what are you talking about?

Nilay Patel (92:51)

They're like, why would.

David Pierce (92:52)

That's not how anything works.

Nilay Patel (92:53)

Yeah, exactly.

David Pierce (92:54)

Anyhow, watch the video. Yeah, no shame. This is like a great video. I thoroughly enjoyed watching every minute of this video. It's so well done, and it is so instructive about how this economy works. And it just. There was just a part of it where I was like. And then within it is the economy.

Nilay Patel (93:10)

Yeah. I say this with deep sadness. Hank Green was really good on Decoder this week talking about a lot of this stuff. And in particular, he talks a lot about why they're able to put up with this game as platforms, because there is this endless supply of people who will come and play the game and be burned out. And once they're burned out, there's a bunch of other people who want to do. Hank talking about everybody will do the silly stuff for six months because it feels good. And there's an endless supply of those people. And I think that is sad, honestly, like, not for those people, but for the. But for the fact that the platforms and the advertisers have decided that they can win this game because that churn gives them all the leverage is a real bummer.

David Pierce (93:54)

And I. And I think there's a real incentive for creators to. To make their bag as quickly as they can because they all think it's fleeting.

Nilay Patel (94:01)

Yeah.

David Pierce (94:01)

It is just a fear that they all express. And again, I'm. I'm saying this with sympathy because that is not a thing that you are in control of if you build your business on someone else's algorithm. And so whenever we post these videos, all these people, like, I get all these comments like, you hate the creators. And I literally, I have done everything in my power to not be beholden to other people's algorithms. Like David can tell you, I'm, like, manic about it every single day that we should be in control of our own platform. And so I have nothing but sympathy for people who have made that choice because I can sense the loss of agency or the precariousness that they feel when there is a shift.

Nilay Patel (94:41)

Yeah.

David Pierce (94:41)

And all I'm saying is go watch the video. I think it's very good, but you can if what, when you watch it through that lens, it takes on a very different tone.

Nilay Patel (94:48)

Yeah, totally. All right, well, that is as good a place to end as any. Because if you want to continue to make sure that Neelai Patel is absolutely ungovernable, the best thing you can do is subscribe to the Verge. Theverge.com Subscribe not only do you get ad free podcasts, but it is like that's the thing that lets us do these things, right? Like, it puts all the incentives in the right place for us. It lets us do the thing that we want to do in the right way that makes sense to everybody and it is working and it makes me tremendously happy. We have gone way over. We should get out of here. It is like 7:30pm thank you as always for watching and listening. This show is a Verge production and part of the Vox Media Podcast Network. This show is produced by by Eric Gomez, Brandon Kiefer and Travis Larchuk. We will be back next week with Alison doing some MWC stuff. We got a lot of other gadget stuff. It's gadget season. We're so back. Nilay Rock and roll.

L'Oreal Group Announcer (95:40)

Support for the show comes from l' Oreal Group. Using the latest advancements in science and tech to create personalized beauty solutions for all, the global beauty leader recently introduced two breakthrough technologies that bring the power of light to hair care and skincare. Light, straight and multistyler and the new LED face mask, both of which were recognized as CES 2026 Innovation Award Honorees. Learn more about both technologies on L'Oreal.com L' Oreal Group Create the beauty that moves the world.