Hello and welcome to the Vergecast, the flagship podcast of The Verge. I'm your friend David Pierce, and right now I am staring at an empty chair, a microphone, and an absolutely gorgeous spiral staircase. And somewhere out of here is going to emerge my co-host, Nilay Patel.
Skipping down the stairs.
I need to enter this show this way every week. What have I been doing?
That was the coolest entrance. Like, you did it too quickly would be my main note for you. Hi Nilay, welcome to the show.
I was so excited.
I'm a little princess.
I'm very sorry to all of our audio listeners that you missed what just happened. But Nilay made a true grand entrance. Imagine me skipping down a spiral staircase.
Mhm. All we need is you and like a long gown and something truly special would have just happened. Uh, where are you? How on earth are you sitting in front of a spiral staircase?
So, I went from San Francisco where I was there for WWDC. I came to LA for a couple days to have some meetings, meet some friends. I'm staying at the W in Beverly Hills. Uh, and it's a very nice hotel. It's been updated, but it's still like an 80s hotel.
And I got upgraded to like the room with the staircase, which was a delight and a surprise. And so, it's very cool. The bed is upstairs in the loft and it still kind of has like 80s vibes all over the place. Like where the TV is over there, uh there's definitely like space for a cable box to be. Do you know what I mean?
It's just very cool. Like I'm in love with this hotel room. I'm like staring at the mountains out the window. It just the collision of trying to modernize an 80s hotel is very apparent in this space and I kind of love it.
This room to me just screams like years of beleaguered actors and now Nilay Patel.
Yes, I think I'm the only person who has never done cocaine in this room.
Still talking about that.
I know what I'm supposed to be doing in here and I'm not doing it.
Yeah, that's very fair. Um, so we have a lot to talk about this week. Uh, there's a bunch of interesting AI news. There's a bunch of interesting social network news. The Hype Desk is back. Brandon Carr is back. We got a whole lightning round coming. Um, but we have to start, Nilay, with I would say one of the most remarkable pieces of news this week. We've talked a lot about WWDC, but a thing has been happening over the last couple of days that we have to talk about, which is that lots of people are starting to get access to the new Siri. And the new Siri by all accounts might actually be very good.
Like it might just be good. I did The Talk Show with John Gruber and Joanna Stern and Joanna had spent that entire day using the new Siri before we did the show and she was just on a tear about the fact that it was good and how impressed she was. I think it might be good. I think there's a chance it's good. My caveat, and don't read this as cynicism. I'm saying there's a chance it might be good. I haven't had a chance to use it. I've been on the road. I'm not updating my like actual phone with a beta while I'm traveling, so I haven't had a chance to dive deep, but I've seen a lot of people use it. I watched Joanna use it.
The thing that is good, the big improvement is that the index of content on your phone actually works now.
Yes. So the bar was you would try to search messages on the iPhone and it would do nothing. And now there's a meaningful index of that stuff, of your mail, your messages, and your photos. That index takes a long time. It actually wasn't even done when I saw Joanna. Like it was still indexing her phone, but even while it was indexing, it could retrieve things. And because there's a useful and good search index on the phone, the AI system, Siri, is able to actually access it, read it, and deliver results, which was the whole promise. I'm not saying any of this like in a cynical or like minimizing way. It's just so funny that Apple neglected to build that thing for so long that now it's I don't even think it's the AI that's doing it. I think it's like functionally they rebuilt the index of content and now an AI can just go get it and that is magic.
Well, it's both, right? I mean, you can't get the index is only as useful as the tool built to access it and like...
Yeah, it's two sides of the same coin. I totally agree with that. I just don't think that there's like some huge leap forward in AI technology happening here.
Oh, I disagree. But I think you're right and you're wrong. Like I think it's very clear that the index piece of this is huge. Like you and I did the episode on Wednesday with the people's biggest questions from WWDC and one of them was what's the thing you're most excited about that isn't all this AI stuff and one of the things I didn't say but had written down in my notes was better search.
Yep. If you've ever used Spotlight on any of your devices it is just an awful search engine. Like if all you're looking for is files, the idea of being able to search for like words inside of a file and it should be able to find it is uncomplicated technology and is something that Apple devices have been bad at for a really long time, right? Mostly at beach balls. Like on a Mac if you try to search your messages, it mostly just beach balls.
Nightmare. Yeah. And so I completely agree that like I think was it Mike Rockwell at the tech talk you went to said they literally rebuilt the thing from the ground up. Like that to me is the single most obvious evidence of the fact that they just rebuilt this thing from nothing. And they started with literally how do we build the database? Because like after nearly 20 years of iOS, they clearly had built the database wrong and they went back and fixed it. So I agree that that's really exciting, but there's also a piece of this that is just Siri's ability to understand what you're talking about has clearly vastly improved. Like anyone who has ever used Siri can tell you that it was bad input and bad output, right? Like it's the reason that you would say remind me to do something and it would like set a timer for 6 and a half days and you're like what just happened?
Yes. Every time I try to set a timer for 15 minutes it sets one for 6 hours consistently for the past 2 years.
Yeah. It's like it was truly like garbage in garbage out. And I think you're right that maybe the most obvious thing that they've fixed is the actual index of stuff on your phone and that raises the floor of what every single bit of AI technology is going to be able to do. But also, it's very clear that they've solved some important pieces of the interface. And I think, frankly, that's where Gemini did the job. Like, Apple continues to be sketchy about what it did with Gemini and the ways in which it's partnering with Google, but like Gemini is an excellent speech-to-text system and an excellent natural language processing system. Like that is a thing Google is very good at. And I suspect Gemini is a big part of how Apple caught up to that really fast because that's a hard thing to build, but it has been built by other companies now and Apple presumably was able to just grab that and can still say it got all that technology without any of the mess that is Google with a straight face.
Yeah, there's some back and forth here and I suspect it will become more clear over time about where the query recognition happens. So I think everyone's assumption is that you talk to Siri, your phone makes a bunch of decisions and then if it needs the power of the cloud and Private Cloud Compute it goes up to the cloud. I heard in some briefings that actually the query goes up and Private Cloud Compute using the full power of Nvidia processors and whatever Gemini distillation Apple framework model thing that they won't talk about in great detail is happening up there. That's making the decision and saying okay you can do this on device and I think it is query to query which one is happening but I think for the more complicated ones that it's happening there and so things like let's search my text messages like find out when this person said this to me I think it's both. I think they have the index on device and then Private Cloud Compute is able to actually go understand what you're saying and ask for the data from the index and I'm very curious what happens if you don't have connectivity.
Or what if for a guy who's been on a plane a lot recently. What if you have really bad connectivity?
I was just going to say these things are even worse when they're crappy than when you're fully offline.
Yeah. And so I think that and that's what I'm saying. I think it will become clear what's happening where as more and more people get access to Siri and they put it under strain in low or no connectivity situations because I think some queries simply won't work. The point I'm making about the index though, just to come back to that, it all depends on what you're comparing it to, right? So, can I search my text messages on an iPhone? The comparison was no, right? The answer was essentially no for the longest time and now the answer is yes in the best possible way. And I think you can get that reaction. What I'm saying about like AI technology is is this an AI system at the absolute frontier of what OpenAI and Anthropic and Google are doing. And I don't think that that's the case. And that's what I mean by Apple hasn't like pushed that really far. Like all of AI right now is like can you find a database? Can you get access to that database? Can the AI go read that database and give you something useful back to you? And even like agent stuff is a lot of where's the database? Can I build you a custom front end to it or like whatever it is. This one is the database was so bad that no one could do anything with it and now it's great and only Apple has access in particular to that one and I do think there's some like magic happening there like it feels amazing.
I just it's just so funny like I'm very focused on well this database didn't even work before now it works and they will never let OpenAI have access to it. That's a big differentiator.
Yeah, I totally agree with that. And I think Allison Johnson on our team wrote a good piece testing some of the I would say sort of second order AI things. Not the like find me this thing in my text message, but like you know compile these pieces of data and find me something. Like she asked what's good for dinner at SFO and it like went and found what terminal she's flying out of because it has access to things in her email and such and gave her some options about places to go. Right? It's like that's step two of AI assistance. In no way is that new, but this is like a thing that you can do in lots of other services right now, but now Siri can do it, too. And it seems to be able to do these things pretty capably. And so this is kind of the big picture question I've been asking myself and others for the last few days is let's just say Siri has caught up to the most basic pretty good AI assistant level. Like it does all of the level one and level two assistant stuff as well as the others do. It's not at the frontier. It's not doing leading edge agentic stuff, but like basic day-to-day life stuff. Let's say Siri has caught up. We have a lot more testing to do before we can actually say that that's the case. It seems to me that if that is true, if Siri is now good enough, this completely upends the entire consumer AI industry for like hundreds of millions or billions of people around the world. If suddenly the thing built into my phone and all of my other Apple devices is good enough, it feels like that suddenly like blows up the universe for lots of these other companies.
Yeah. I mean, this is why I keep saying Apple sherlocked the free version of ChatGPT.
Yeah. Yeah. The free version of ChatGPT was a sensation because Google search had become so and Siri was useless and now here was a thing on your phone that would just talk to you and maybe just maybe lie to you. Maybe maybe try to break you up with your wife in the front page of the New York Times. Shout out to our boy Kevin Roose. Um, but at the end of the day, that didn't improve enough, right? OpenAI's goal is to get you to pay money. And they even, you know, the sort of like false start with Apple Intelligence, OpenAI's strategy was still to give it away to Apple for free and hope that you would use it enough to get you to pay money to OpenAI for the good version of ChatGPT. And none of that ever worked out. And now OpenAI is pivoting to enterprise and they're all chasing that money. And I think Apple is just saying, look, this is good enough. This is maybe as good as two years ago free ChatGPT. This is essentially the same level of capability, but because it has access to your iMessage and your photos and if you use Apple Mail, your mail, a lot more can be done here. And we're building visual intelligence directly into the camera, which is a huge deal for iPhone users.
That's a lot. There are straightforwardly that's a lot to build into the operating system. What is the thing you're going to have to sell to get people to pay money or download a different app is a very complicated question when that's built into Siri. And I really think like the free AI apps are all kind of they're all going to put out statements soon like we welcome the competition, you know, like that's the thing that happens when apps get sherlocked in this way and it is just here. It's absolutely going to be the thing that happens now.
Yeah. I've been thinking a lot about this being sort of the equivalent of like the can Netflix become HBO before HBO can become Netflix thing. It was like you and I have spent a lot of time on this show over the last couple of years talking about all of the AI companies trying desperately to figure out how to be a platform, right? Like ChatGPT has taken every imaginable run at trying to build app stores. Everybody is doing skills. Everybody is building connectors. MCP is a big part of trying to stitch this whole system together. There's this question of like how do we take a chatbot which clearly people like interacting with and break it out into something bigger and broader and sort of cross app and more functional. How do we give this tools and things to do? And the bet was that we can get there before Apple can figure out how to make a pretty good assistant because it actually has all this other stuff. And Apple, even more than Google, has that kind of access and sway over developers to get them to play some of these games. So, I think good Siri is going to have a pretty easy time getting most developers to do things like expose app intents and give Siri access to their tools in order to do stuff. I know you're going to DoorDash problem me in a minute, but it does seem like Apple is now on its way to winning that race to becoming ChatGPT or at least you know free ChatGPT before ChatGPT could figure out how to be the app store and that to me if I'm OpenAI is like a very scary state of affairs.
This is why OpenAI is pivoting to enterprise. I think they have realized that there's not money in the consumer market. I'm the one who keeps yelling that there's no great consumer AI products, right? Like, there hasn't been the thing that would make people pick up a different device or have a different experience before picking up an iPhone. And if you don't have Candy Crush and Instagram and TikTok and YouTube, what do you got? You know, like your phone is going to be with you because those are the things that consumers want on their phones. Your weird Meta glasses aren't going to deliver all those experiences in that way. Certainly your friend pendant which is essentially useless is not going to deliver those experiences in that way. So either you lean all the way into people don't even want apps anymore. They just want to be free. They want to be talking to their necklace. Whatever you think that future is or you have to disrupt the phone. And you know OpenAI has Jony Ive. They're actively going to try to disrupt the phone with a phone. We think maybe with their own phone. Um, you know, Microsoft is like showing off like lanyards. Like they've just got weird ideas about what comes next. And I think Apple, they're like, "Well, all of your products are apps on our phone, so we can just do the thing that we're really good at, which is disrupting the apps on our phone by building the features in the operating system, and then whatever comes next, like we'll be ready for it." The big question for Apple is they're so device centric for privacy reasons, yes, but also that's the app model, right? All the apps live on your phone. They don't live in the cloud. So yes, they can do agentic stuff and they can call you a DoorDash. What are you going to do when you just have AirPods in or you have Apple's glasses in and you want to call an app? Is that going to call your phone in your pocket? Are you going to leave your phone at home like you can with an Apple Watch now? It's really unclear where that app logic lives beyond just the I can strongarm app developers into doing what I want. Right now it's like okay, we're going to go and use the web version of DoorDash. Google is much better at that. Like straightforwardly, Google is much better at that. And their strategy is we will open a Chrome browser on Google Cloud for you and click around DoorDash and they can just do that. Apple doesn't have those moves and they're still very focused on selling devices, running the apps locally in devices. Even the index I'm talking about, Gruber pointed out to me, they rebuild that index per device. So when you update to one of the 27 OSes, your Mac builds its own index, your phone builds it, your iPad builds its own index. They don't share it. So, this device model is going to be freeing in one way because they have access to your data, limiting in another really big way.
Yeah, that's interesting. That actually makes sense. I was running the new Siri on my iPad, which like I don't use for things like calendar. So, I asked it a calendar question and it turns out I've never opened the calendar app on this thing. So, it had no idea. I was like, "Oh, well, this totally failed." And then it was like, "No, I just literally have never actually looked at the calendar on this device." Um, but I agree with what you're saying and I think I'm curious how you feel about the way Apple has been talking about that stuff because um at that same tech talk I think it was Craig Federighi alluded to basically Apple thinks that a lot of people don't want to use their AI systems the way that software developers do. Uh and I think that is very clearly something you and I agree on. Um, but this question of is Apple basically saying that we don't think these sort of multi-step orchestrating agents is the future of consumer technology. Are they saying that because they haven't built it yet in the same way that they pretended that AI wasn't important because they screwed up AI so bad? Uh, or is this actually like a statement of UX from Apple that they're like, "No, your job is not to orchestrate agents. We just built a thing that can be helpful." Right? Like they even talked about it as we don't think of Siri as sort of the center of the universe. They almost had to sort of answer for why they built an app at all. And they basically see it as a way to go see old conversations, not like the place you go to use your phone. But they see it as sort of across the systems you already use. And sure, if you're Apple and you have all of the incentives Apple has, that's a thing to do. But I also think it's not a crazy case to make as a product person to say actually most people don't and won't think of computers as a bunch of agents to send to do things on their behalf. They just want to order food.
Yeah. What was the title of our last episode that we did together? Where's the computer? This is a central question in the age of AI. Where does the logic live? Where does the computing happen? And if you want a bunch of agents to go do stuff for you, it should be the cloud. The answer somewhat naturally is the cloud.
Because then you can just issue commands and you don't have to worry about your phone battery dying or whatever it is. The things happen and then the result is delivered to you. I'm less of the opinion that consumers care or don't care about multi-step agent orchestration. Like sure, they also don't care about Docker or HTML or like what you know. It's like yes, like technology exists. Our audience cares a lot that Bluetooth 5.0 exists. Sure.
Do normal people understand that Bluetooth 5.0 is what enables a bunch of wireless earbuds to have pairing across the world? No, of course they don't.
So, I think Apple... Damn it, uh, we're going to do a great holiday spectacular this year. Like, I think there's a little bit of difference between the tech companies, the sort of enthusiast audience, which we have here, and then like the mainstream technology consumer, right? And somewhere in there, Apple's pendulum swings back and forth.
Like when it's time for a new iPhone, we are going to hear exactly the details of the new A-series chip. If the iPhone folds, we will hear exactly the details of how they remove the crease from that OLED screen. They care about the technical innovation when it's theirs. When it's someone else's, they're like, "No one cares about this." Like, "No one has to care about any of this stuff." And it's like, well, the stuff is what you make the products out of. I don't think consumers need to care about multi-step agent orchestration, but if you do want to say, hey, I'm landing in LA in the next hour. Make sure I have an Uber. My hotel is checked into and there's food waiting for me when I get there. Well, yeah, that's a pretty normal consumer use case. It's a pretty normal use case for anybody who's traveling. You need multi-step agent orchestration. Like it doesn't matter if you know that that's the thing that's happening or not.
And that's the gap that I think Apple's in in the spot where they're saying no one cares about the technology and what would you even use it for? You're just spaghettting it at the wall which is what they do while they let the rest of the tech industry figure it out. But it's not to say the technology is not important.
Yeah, I agree with that. Thank you. Next time we do a Vergecast movie night, we need to do Her because Her, a shockingly prescient 2013 movie, I think, has really strong opinions about how an assistant might tell you when to look at a screen uh in a way that I think is really fascinating. But like that to me is I think I don't know. I think the part of what you were saying before that I keep thinking about is uh this idea that I'm just going to be on AirPods and that I might order dinner without any other device ever but AirPods. That I just sort of declare to the world through my AirPods that I would like pizza and it does all of the work and pizza arrives at my house. And I think Apple might believe and it might be right that that's just not how people are ever going to use their devices. And that Apple actually has a huge advantage by being able to be like, "Hey, open up your phone to review your order just to make sure we're getting it right." That like that's the step before multi-agent orchestration, right? That is like this is the human in the loop stuff we keep talking about that uh how do we make this work in a way that feels like normal people using normal devices in normal ways and I think again it's very possible that all of this is just Apple continuing to be behind on technology and trying to invent reasons that it's behind on purpose but I do think there is a real case for Apple coming out here and being like Siri is not the immediate way that you are going to do most things on your phone from now until the end of time. But it is just another feature of how you already use your phone.
Like I sort of hope that's the answer.
Apple's addicted to weird riffs on standards that no one else uses.
You know, you could probably build all this stuff with MCP today. Is Apple going to have a weird riff on MCP that it's going to force the entire industry? Of course it is. Absolutely.
Just like it has a weird riff on Dolby Vision that the entire industry has to use to get Dolby Vision on an Apple TV. Like it's just the way it goes. Uh and that's how they maintain control of their ecosystem. I think this is one of the reasons that the European Union is like no like you have to let things be interoperable. If your AI assistant is the only thing that works at the operating system level in this way, there will not be competition for AI assistants and we want competition this time around. You know, Apple's response is to say screw you, we're going home. Maybe that means more people use Claude in this way or Gemini in this way in Europe and there actually develops a parallel ecosystem or maybe Apple figures out whatever technical solution it currently says is too hard to figure out or that it can't get approval for from the EU and Gemini or ChatGPT or Claude get to tell you, hey, look at your screen for the human in the loop verification. And that power of the ecosystem right now is all tilted in Apple's favor. That's the argument you're making. I think you can see Europe saying actually we need to make sure we break it open this time so it's not just a monopoly or a duopoly.
Actually, let me ask you this. There's a big argument to be made based on all the Siri reactions that once again iMessage lock-in is the thing, right? All these texts are in iMessage. No one else can see them.
Boy, Siri is great now. I mean, if you were a person who like a sicko has had never delete as the option in your iMessage and so you just have however many years of messages, you are in for a wild surprise when that index finishes with the new Siri. Uh, but yeah, I kind of think you're right. I mean, there's the lock-in certainly is going to be vastly more intense now because the switching costs of losing all of those messages even if you decide to move, right? Like if I switch to Android, I've now done this. Like switching to Android is easier than it used to be. But the idea of like my assistant is now going to have vastly less stuff to work with because there's a huge source of my life in those messages is a big deal. Like the lock-in feels stronger than ever if Siri is good because good Siri is going to be very hard to walk away from and to start over from with one of these other things. Like, and again, this is something all of the other AI services have been trying to do, too. They all have memory. They're all like asking for custom instructions, right? Like, the idea is the better we know you, the more we know about you, the harder it is going to be to leave because you're going to have to start over with something else. And the fact that, like you said, Apple fixed this database just gave it a huge advantage on that front that is going to be very hard to walk away from for a lot of people.
I feel like if two years ago they had said, "Here's what we want Siri to do. We're gonna spend two years figuring out how to search messages." I think a lot of people would be like, "Yeah, that's how long it's going to take."
Yeah. And then we're going to be able to find all of the stuff in your messages that you can't. Now, everybody would be like, "Great. See you in 2026. Can't wait."
Yeah. Yeah. I'm fascinated by this. I need to test it so much more, but like the number of deeply infuriating basic tasks that Siri can't do seems to have gone way down. And just that alone, what a great day this is. Like if it just remembers my reminders correctly from now on, this will have all been worth it. AI is worth it just for that. Um all right, we just take a break and then we're going to come back and we have some social networking figuring out to do. We'll be right back.
Support for the show comes from Anthropic. Important questions don't usually have simple answers and they usually come with a mix of uncertainty, excitement, and reflection before a big breakthrough. So, when you're working through something important, it helps to have a sounding board to explore ideas and uncover what's really underneath them. That's where Claude comes in. Claude is the AI for minds that don't stop at good enough. It's the collaborator that actually understands your entire workflow and thinks with you. Whether you're debugging code at midnight or strategizing your next business move, Claude extends your thinking to tackle the problems that matter. Plus, Claude's research capabilities go deeper than basic web search. It can have comprehensive, reliable analysis with proper citations, turning hours of research into minutes. For problems worth solving, get started with Claude today at claude.ai/verge. That's claude.ai/verge. And check out Claude Pro, which includes access to all the features mentioned in today's episode. claude.ai/verge.
Support for the show comes from Granola. It feels good to walk away from a meeting that actually felt productive, but our brains can hold only so much info. You might look back and realize you either A got caught up in your notes and missed some important stuff, or B, you couldn't even take notes and now you're just trying to reconstruct it all from memory. If that's your dilemma, Granola can help solve it. Granola is an AI-powered notepad built for the way real people actually meet. You can take rough notes like you normally would while Granola securely transcribes the meeting. Then after you wrap, it turns everything into clean, structured, actually useful notes. Granola also works through your device's audio, which means it integrates seamlessly into the video conferencing tools you already use. It's your standard meeting setup, but enhanced. If meetings are eating up your day, Granola is a no-brainer. You could try it totally free for 3 months. Just head to granola.ai/verge. That's granola.ai/verge to get your time back. Get three months free at granola.ai/verge.
All right, we're back. Um, Nilay, I want to give you three small pieces of news and then I want you to galaxy brain your way through them for me. You think we can do this?
Okay, here are the three pieces of news. Piece number one, Instagram is letting you more explicitly define your algorithm in Instagram. You are going to be able to tell Instagram what you want to see more of in your main feed. Piece number two, Bluesky is getting a new feature called communities, which is basically a smaller version of Bluesky for people who want to talk to each other. It's Reddit but on Bluesky. Piece number three. YouTube has introduced DMs for somewhere between the second and 53rd time in the history of YouTube. There's something going on in social networks right now and all three of these things feel like they are part of some turn in how we think about social media and social networks. Do you have a grand theory of this moment in social?
Can I add a fourth thing that is utterly self-serving, but I think fascinating in this context. We have a really great product manager named Danielle. If you follow our site, you can see she does like version updates like every week. Like here's what we shipped this week. She's great. She does a lot of user research. And when we were doing our homepage refresh, she did a lot of user research about the story stream and quick and all that stuff. And one thing, the thing that I'm adding to this piece number four is a lot of people tell Danielle, like a lot of Verge readers tell Danielle, they use our story stream instead of social networking products, which is always our goal. So, this is very self-serving. Like, I'm very proud of this thing, but they're like, I'm tired of algorithmic social media. Here's a thing that feels like that with a community I like. Great. We're nowhere near the scale of these other networks, but I'm just adding that to the mix because it's something that our audience, maybe some of you listening to this have said to us. We're tired of big social networks. And all of the things you're describing are ways to make the big social networks feel small. And I think that's fascinating.
Uh let's start with Mosseri. Uh because I could talk about Adam Mosseri's Threads post about the recommendations algorithm and the sense of agency and control all day and all night. And I've asked Adam to come on Decoder or The Vergecast to come and talk about this because I could spend hours and hours talking to him about this post and what he thinks agency and control means in the context of a giant multi-billion user algorithmic social networking product. But here's what he wrote. Something is shifting in what's technically possible. For years, ranking algorithms have been built with technologies that aren't legible to people. No human can read a neural net and explain why an algorithm thought you might be interested in a given video. You can't have agency over a system that lacks an interface you understand. And then he goes on to say, "What's new is that LLMs can now look at clusters of content and describe them in language people understand." Right? So all of this is just math. And so any given video, the algorithm was just putting together like coordinates to say inside of our giant database of content, these all match in a cluster. And you could never look at that. It would be like looking at the matrix. And now the LLM can look at that at scale in low latency and say this is why this video is here which means you can talk back to it and you can say I want more of this and less of that. So what he is describing is not just like a set of filters. It's not just you know ask it for more puppy videos. It's like very directly an interface to the algorithm itself that you can now talk to in natural language. And his point is, I'm doing this because I want you to feel control. Right? Recommendations are great. Not all of your friends are posting all the time. If you open Instagram and it's just the last five posts from your friends that you saw the last time, you won't be as happy as if we show you something else interesting that's happening in the network. You know, people like viral content for a reason. Now, we can let you talk directly to the algorithm. We want you to feel a sense of control of what's happening. That's what the LLM is for. That is a way of making it feel small, right? Making it feel in your control. Again, I could talk about this all day and all night. I'm not sure this matches how Meta makes money by showing you ads.
Right. Across recommended content. I'm not sure it matches Meta's entire approach to engagement maximization or the fact that Threads, which is built on all the same algorithms, appears to be built just to make dumb people as angry as possible.
It is engagement bait as a platform. Like it really, really is.
Like Threads in particular is that thing, right? Yeah. Uh, so there's a lot here, but you can see I mean Mosseri's argument for why they built this is not just we can. It's like there's a lengthy reason to post for it and it's to make it feel small.
Bluesky Communities, it's funny. Everyone thinks Bluesky is a platform like it's a product. Like I've talked to the new interim of Bluesky, Tony Schneider. They are and they have always been a protocol company, right? They're trying to build the next generation of decentralized social protocols. Mhm. And I think what they see as communities here is they would just describe it as...
Like schema. Like how do you describe a New York Knicks community? Like here's the group of people talking about the Knicks in a way that is legible across the whole atmosphere so that if you make a Reddit on the atmosphere that works in open social, you can say, 'Oh, there's a community of Knicks fans on Blue Sky. I'm going to show that community in this app in a way that is coherent.' And that's a big deal, right? It's we want big thing. We want little groups of communities to travel in different ways across the big network. That's a way of making the big network which up until now is like here's a fire hose of content. You can filter it. Now it's not just a fire hose of content like here's a group. Here's a taxonomy of what a community is which is by the way you and I have talked about this many many times over the years. Describing what a community is online is very challenging.
Yes. Like is book talk a community? What are the boundaries of book talk? Is it moderated? Is it controlled? No, it's just a bunch of people who say they're book talk. That's it. Yeah, that's that's all it is. Uh there are platforms like early Twitter communities were just people using the same hashtag. Is that a community?
Reddit has a very different model of what a community is, right? People start them. They are in control of them. The moderators do stuff. Reddit as a platform occasionally goes to war with its communities. People split off and they they have rival subreddits when they're mad at the moderators. totally different model of community all to make a big platform feel small in different ways. So I think this is Blue Sky saying, 'Okay, the the protocol itself has to reflect the fact that communities form here because you can't just rely on hashtags or people declaring they're part of book talk.' And so you see again, it's another attempt to make a thing feel small. And I think it it truly with YouTube, they have to admit that they're a social network every now and again. And so if you do shorts and you have people in the comments, it is only natural those people should want to talk to each other, right? And particularly the number of creators who are like comment with whatever, you know, comment with course and I'll DM you a discount code. Like that is actually a monetization funnel for a lot of like weird creators.
Yeah, I think that the DMs one is is kind of the most straightforward to me because everything we've seen from YouTube over the last year or so has been about keeping you on YouTube for more activities longer. And one obvious one is uh let people send videos to each other, which is just like not really a behavior that exists inside of YouTube. And so you can build all of these other monetization systems into YouTube. You can build all of these new like super subscription things into YouTube. They're they're just trying to make it so that your entire experience of YouTube can stay inside of YouTube. Which is very funny because if you remember way way way back, one of the main ways YouTube became the default video player on the internet was because it had really great embed that you could put anywhere on the internet. Like you YouTube's embed system was better than everybody's for a really long time and it was like a thing the company took great pride in and they've actually kind of made it worse. like YouTube embeds on our website right now are worse than they were a couple of years ago and it is because YouTube is really really really trying to get you to go to youtube.com and so just DMs is just a way to send videos to each other strikes me as like an obvious answer to that.
And and by the way Instagram knows this. Adami says this all the time that the the number one mechanic on Instagram is you watch reals and send it to a friend. YouTube doesn't have that mechanic right like it it just never has and I think part of the reason this is Again, my point about trying to make everything feel smaller. I don't perceive that my friends are on YouTube.
That is not a social product in the same way.
Right. Like I open YouTube to get something done and I I leave. I know a lot of people watch YouTube all day and all night, but the idea that that's a social experience you're having. This is what I mean about they have to admit that it's a social product. Yeah.
And so they have to just capitulate to on social products, people need to be able to message each other. And that mechanic of you watched a great reel and you sent it to your friend and then you have a little conversation and now your engagement time has gone up and we can surface you more ads. All that requires just admitting that reals has made YouTube vastly more of a social network than traditional YouTube ever really was.
Yeah. I was I was listening to a podcast the other day where they were they were talking about how male friendship in 2026 is just sending each other Tik Toks. Uh I was like that's super real, dude. like and and YouTube obviously wants to pee that.
Oh, dude. The that the the the my Christian Dior meme. I mean, that's brought that's brought more men together in the past
three weeks than like maybe anything in modern society.
Like that thing is going to save America.
It's really true. That and the random New Zealand soccer player. Have you heard about this?
So, there's an Argentinian influencer. I'll I'll link this somewhere in the show notes, but there's an Argentinian influencer who went and found the person they deemed to be the most random sort of unfollowed person at the World Cup. And they identified a a New Zealand soccer player named Tim Payne. And we're just like, let's blow up Tim Payne. And he went from having 4,000 followers on Instagram to last I checked it's well over 5 million. That's amazing. Um, and he has just become he is now like a phenomenon of soccer just in the last like two weeks before the World Cup. And it is like this is another person like Tim Payne comes up in every group chat that I'm in now. It's great.
Sounds great. But let's go back to Blue Sky for a minute because uh this community's announcement comes uh what a week maybe after uh Blue Sky COO Rose Wang did an interview with CNBC where she kind of teased this happening uh basically saying that they were more inspired by Reddit going forward than by X or threads. But she said this thing that I I've been thinking about ever since, which is um she said, 'What we've learned through this process is that I think the public square is not the direction we want to go in. Essentially, I think it's useful as a discovery mechanism, but we're very inspired by companies like Reddit. A public square where there's only a stage and there's posters, like people on a stage and people who are watching, that is not social. We're in the medieval stages of the online world.' A completely agree that is like fundamentally not a social product. But B, I think I think this communities thing might be a bigger pivot for Blue Sky than you're giving it credit for. That like this thing popped up as this is supposed to be Twitter, right? Like Twitter is dead. Blue Sky is the new place to go. Everything's going to be fine. Threads did the same thing. They like they built threads really fast in large part to capitalize on this idea of people wanted to leave Twitter but didn't have anywhere to go. Neither of those has worked in that way. Neither of them has become the new Twitter. Blue Sky grew really fast and has has really kind of tapered off in some interesting ways. Threads as I don't know that we've heard numbers in a minute, but presumably continues to grow, but is like it's not what Twitter was.
Threads is growing because they if you open Instagram, they're like, 'Look at this threads post that's also a picture.' And then you click on it and it opens threads.
Yeah, threads will never be what Twitter was for a very long time. But but I what I wonder for Blues Sky is like do you look at this and say, 'Oh, actually maybe what the end of Twitter was was the end of this kind of era of social period and that Instagram got away from that. Facebook is no longer that.' Not only is Twitter not that thing, that thing might just be a dead activity of like giant primarily textbased posting systems where everybody has a a single feed in sort of real time ways in the sense that like maybe maybe that is gone and that everything has moved to these smaller communities and group chats and and private sharing and all of this stuff. And so like one way to look at this, I agree, is that Blue Sky has always attempted to be more than just the Twitter clone and the atmosphere and the the app protocol is a bigger deal than Blue Sky. But this also kind of feels to me like a complete pivot for Blue Sky away from the idea of being Twitter at all.
You know, the former CEO of Blue Sky, Jay Graver, she's still at Blue Sky. She's working on sort of the technical aspects of it now. and Tony's the interim CEO while they figure out what they're going to do next. Uh Jay was on to coder ages ago and what she said to me then was we made a thing that looked like Twitter because one they came out of Twitter. If you will recall blue sky was a project inside of Twitter that got spun out.
But the reason they made a thing that looked like Twitter is they just needed a product that people could understand while they worked on the protocol that they were actually interested in. And this is back when Jack Dorsey, you know, would would pronounce that Twitter never wanted to be a company. It really wanted to be a protocol. And no one understood what he meant. And I think Lucy is a real extension of that. What you want is an open social protocol. And if that's the thing you want, then you will never build a product that is great. And you can I think you can see that with Blue Sky today. There's a reason that sure, you know, Blue Sky has whatever reputation it has, but they're not actively trying to go change that reputation or recruit users or build a bunch of features that people love. All of their time is on the protocol. And I think what they've discovered is, yeah, I I agree with you. The idea of the big influential Twitter may have just come to an end. And Twitter itself, you know, we can see in the SpaceX IPO, Elon has destroyed that business. The users are down, the revenue is down. The only revenue that's up at at X the everything app is data licensing to XAI. So like in a in a circular way.
Yeah. And I'm certainly doing all of my credit card processing in X. I don't know about you.
Who doesn't want to trust their financial data to Elon Musk? So you see like maybe that era has come to an end and maybe threads will never actually replace Twitter as the the I don't know. But Blue Sky is saying the next generation of all these platforms should be interoperable. So if you want to build your own version of Twitter and you want to go for it, you can accept this fire hose of content from the atmosphere and build something out of it. And if you think that Reddit is dumb and you want to build your own Reddit, you should be able to look at communities somewhere else on the web and surface and display those communities in ways that are better than who else whoever else would do it. The the problem for all this and I think everyone knows I'm like a believer. I believe in this. The problem here is that there is no shipping validation of these ideas anywhere.
Because threads is on activity pub and blue sky is on app protocol and yes there are like bridges and whatever to make them you know fake interoperate but they don't actually interoperate. So the two big theoretically interoperable social networks are in their own silos and there's there's literally no other thing that's big that proves the point for either of them. Like Masttodon does not prove the point for threads. Like I'm very sorry. It just does not. There isn't another big atmosphere product that proves the point for Blue Sky. And there's a lot of interesting ones. Pixel Fed is interesting. Skylight the video player is interesting. they do not have the scale to prove the point that they're interoperable. I think maybe until they interoperate with each other, that thing will not exist. Or if someone sends up a Reddit, that thing will not exist. But I do think Blue Sky thesis is that the internet should be more open and it's not their job to build all of the products. I think that confuses everyone all the time.
I I really think everyone thinks Blue Sky is a product company and they are just not.
Yeah. But I think like one of the things that we've always struggled with in explaining why we're so excited about the Fedverse is like explaining what it looks like, right? And I think for a long time the running theory was that the sort of base experience of it should probably look like Twitter that it's it's a bunch of posts in a row and that the cool thing about the Fedverse is you can you can reconstitute that in whatever way you want to, right? that what you have is a giant bucket of content and you can write into that bucket and you can read out of that bucket in whatever way you want and that is cool and exciting in theory but that probably what most people want and the way most people will experience it is something that looks like Twitter and I think that might have been right four years ago and I I think it seems to me that everyone is rapidly coming to the conclusion that that is wrong and that that even as you look at something like the algorithm changes on Instagram like to your point it's about making it feel smaller, right? Like this is this is me saying there's this giant corpus of content. I want all the gardening stuff. That is not Twitter at all. That's Pinterest. Like that's that is such a fundamentally different way of looking at content than this idea of rapidly moving reverse crown posts. And I think it's very possible that we are just at the end of that era of the internet. And I just think that's fascinating.
Yeah, I agree with that. And I I do think maybe this pivot to communities is a response to that. I I think all I'm saying is it's not I don't think Blue Sky is launching communities. I think Blue Sky is launching a set of extensions to app protocol that support communities. Do you know and it's like just like very different like that company is very different in how it thinks about what it makes. And I I do think that gets everyone confused all the time because people talk about Blue Sky as though it's a competitor to Twitter and I don't think it has any desire to be that thing.
Yeah, I agree. I also just went to Blue Sky and it just popped up a thing and it said new feature group chats.
Yeah, this is just we're all doing the same thing again here.
Um the algorithm piece just let's just go back to Instagram for a minute here cuz I I think this is also really interesting. Like there's this broader trend happening and Tik Tok has done a little bit of this and uh I think Instagram has been poking at this for a while. just the idea that you should be able to have some more proactive control of your algorithm that like threads has the thing where you can literally say dear Algo uh and Tik Tok lets you sort of tell things tell it things not just that you're not interested but that you are interested in. I've been trying to figure out to what extent this is actually like a real change in how these companies run their products versus sort of a placebo to make it feel like you have some control where actually all it is is continuing to measure your behavior and showing you videos that statistically you're going to watch even if they make you feel gross and bad. I love the idea that we can give everyone more proactive control over their algorithm, but like you said earlier, I also wonder if that is just frankly bad for business for all these companies. All right, you've walked into the end of the Miseri post, which I've been saving for last.
I confessed it's really long.
It's very long and everyone hit the heart of it, which is you can talk to your algorithm now. Yeah.
Okay, here's the end of this.
Yeah, this is adamissary. This is the start of something bigger than a feature. It's in our best interest as a business to empower people to shape Instagram into something that works for them and that people should have a meaningful amount of agency over the products they spend so much time in. We intend to build much of what comes next on that principle. So that's great statement of purpose agreeing what you say.
And then he says there's a harder version of this question on the horizon that's worth saying out loud. Within a few years AI will be capable of not only letting us see and shape algorithms but also generating entire bespoke experiences on the fly tailored to an individual in real time.
At that point you can imagine shaping much more than ranking. You can shape the structure of the app itself. The experiences inside it. Even the things the app is for could be different for each of us. Okay. Okay. And then this is here's the Nei bait. Like I read this and I literally opened my email and I wrote to Adam in like all capital letters like come on to coder and talk about this. This is a real thing. Um that's a real thing that I did because this is what he wrote. A world where you are making personalized experiences on the fly is exciting in terms of agency but at the extreme starts to undermine shared experiences. If AI can generate entire apps and experiences that each of us wants, you and I might no longer share any sense of space. Wow. Right? That's like the biggest idea. Like you have an Instagram, it's like not only different content for you, not just a different algorithm, but a whole different app that is doing different things for you,
filled with different kinds of synthetic content.
Different synthetic content, different buttons,
right? Like the Instagram you might see as somebody who needs to sell ads for your lawn mowing business might be totally different than the Instagram someone might see who just wants to see get ready with me. It's like
a different app. Like they will be different apps with different content and like different business goals for Instagram. If you recall um when I interviewed Soar this last time I said look you're going to ship different apps to everybody based on their search results. Like you're going to undermine Google as a shared sense of truth and his response was it's a spectrum. So, if you ask what is the capital of California, you'll get an answer and that answer should be the same for everybody. But if you ask something like what's the best place for me to go in California, Google will like build you a trip planner, right? And he's like, that's a spectrum and we'll just figure it out. And it's like, oh, there's a whole blurry middle of that spectrum. It's actually pretty blurry, man. Uh, what is the body of water between Florida and Texas called?
Yeah, there's one guy who definitely wants a different answer than I do. Uh, the Tap Bridge. You're standing out front of the capital. What happened here on January 6th, 2021 is forever the great example.
Uh yeah, was the election stolen? Who knows? Um I know the answer, by the way. It was not. Um but I know some people who are very invested in Google issuing a different answer to that question. Okay, now bring that to Instagram and Instagram is rewriting itself on the fly to show you different content and make different experiences for you and you can direct it in natural language. Like Adam is pointing right at it. He's not shying away from the messy spectrum. And I I you know I you can have a lot of feelings about Meta. You can have a lot of feelings about Mark Zuckerberg. You know a lot of feelings about Adami, but it is clear in at least the times I've talked to him that he has been the most thoughtful about this. Like he sees all the problems. You might disagree with all the trade-offs he's made and all the decisions he's made, but he's been the most like, 'Yeah, uh we're going to screw everything up.' Like he does see the thing. Yeah, he sees the thing very clearly and he's like, 'Yeah, whatever trade-offs we're making, we're making them. But at least I'm copying to the trade-offs. I think that's fascinating. Again, Adam, you know, I I will we put up the screenshot of like five exclamation points that I sent to him. Um uh I've talked to him many times over the years. I would love to spend a lot of time digging into this because this is I think this is the future of computing in a real way, right? Is these like bespoke experiences and it will undermine our shared sense of truth, our shared sense of experience. You know how people say like Reddit is mad about something that implies a structure and experience of Reddit that allows that to mean something. If you destabilize that, it doesn't mean anything at all. Right. And there there's something there that's important.
Yeah. We're at this fascinating moment where we we totally sort of destroyed the idea of a monoculture, right? The idea that like you and I have shared cultural references just increasingly has gone out the window. Everybody has different algorithms. Everybody has different apps. We see different stuff. But there's still a sense that some other people have seen the things that I'm seeing, right? And all of these communities now exist inside of these apps. So that like I can't trust that you and I have seen the same things, but I can I can I I can identify with the other people who have. It does there there's a limit to how bespoke you can get before we're all just hanging out by ourselves. And I have to believe like for my own sanity, I have to believe that is not an experience that people want.
I don't know. I don't either. And like and I think there is a real if from a sort of ruthlessly efficient business standpoint that probably works for Meta to some extent to give everybody exactly what they want without any regard for its connection to the broader culture or society might work. We we can just spin up the exact thing that you're looking for right now. No one will ever see it but you in this moment. There are a lot of reasons that's really good business if you're meta. And boy is that just like a bleak bleak universe to live in.
I mean that is Mark Zuckerberg's vision of the metaverse in a very real way. Like all the idea has been the same regardless of whether you're wearing a helmet or AI glasses.
I think the metaverse is less bleak. At least in his world there are other legless people with you in the metaverse,
right? But your your brain's still in a vat and he has total control of you and like you know what I mean? like he's like instead of buying a TV you'll buy my TV which displays my content and we'll just like generate it virtually like his version of this is that you live in a synthetic environment under his control and again whether that is happening in a headset or in glasses or just on Instagram that's it's the thing that is being pushed towards like every pixel will be under the control of meta and it I think it is just fascinating to see adversary just come right out and say there might be some problems with that
weird. It just doesn't You don't hear many of these CEOs ever just say that out loud.
Yeah, this new feature might be the end of shared human experiences. We'll see how that goes.
Like we're turning this knob. Uh let us know when it destroys the world. Like it's just you never see it. I read this and I was like I again I could talk about all day and all night because it it's so self-aware that something about all of us using our phones at the same time actually did create shared experience. As much as it has driven us apart, there's something very important about the structures being the same.
Yeah, it's good. Adam, come on the show.
This show decoder, which I don't know. I I get AI summaries.
We'll do we'll do Subway takes together, man. Whatever. I just want to do the whole thing. I'm so You can tell I'm so into it.
We're going to do a chicken shop date, but it's Nei and Adam Miser.
I would do hot ones with Adam Miseri.
Sold. All right, we're going to take a break. Then we're going to come back. It's lightning round time. Be right back.
Support for this show comes from Clavio. There are only so many hours in a day, and Claio's two powerful AI agents can make sure your team spends them on big things. The first Clavio AI agent turns your marketing ideas into reality instantly. Describe what you want. A holiday campaign, a VIP re-engagement series, and Clavio builds it instantly. Email, SMS, and push. All coordinated on brand, grounded in 14 years of Claio marketing data. Nothing goes live without your say so. The other Claio AI agent keeps your customers happy at any hour. Brand trained to answer questions, make product recommendations, and handle orders and returns. No hold music. Marketing that launches instantly. Support that never sleeps. Join more than 193,000 brands, including Away, Patrick Ta, and Dollar Shave Club. Already growing with Clavio, the autonomous BTOC CRM. Get started at kl Avi yo.com.
All right, we're back. It's time now for the hype desk when our friends Ross and Ashley come and tell us about what's cool in the world this week. No Ross. Screw you, Ross. Just Ashley Essketo. Welcome back, Ashley. I'm so You finally fired him. I'm so excited.
Honestly, it was long overdue. This is like the third time for me. Honestly, this is the longest game I've been playing. Let me tell you, I had to go into business with this man.
Uh, what have you brought for us this week?
Oh, boy. Guys, we got to talk about the social reckoning trailer because you know what's cooler than a million dollars?
What? Whatever not this trailer is. I don't know. I'm uh I have a lot of questions. A lot of question. This is this is the the sequel to The Social Network. The Aaron Sorcin, Jesse Eisenberg, Dawn of Facebook movie.
I will say, Neili, you missed a very important name in those names, which is David Fincher, the legendary director who directed that movie.
Yeah. And made it very stylish. Made it very stylish.
He did make it very stylish. Also, like Trent Resnner, Anakus Ross did the music.
That movie had a lot of moves going on. A perfect storm. a young Dakota Johnson making a a spectacular cameo appearance.
You cannot force lightning in a bottle, but that's about as close as you can get to doing it. You know what I mean? Like if you Those are the ingredients and maybe you get a good reaction like that that was those are the ingredients. But this time it's just the Aaron Sorcin show. He has written this. he has directed this. And um and now uh the eldest boy, everyone's favorite eldest boy. Uh Jeremy Strong, who played Kendall Roy on Succession, is playing an older Mark Zuckerberg, which is odd because wasn't he like 35 or 36 when this happened? Like Zuckerberg and then but Jeremy Strong is like 46. So he adds a little bit too I feel he adds maybe too much gravitas. Zuckerberg, not a person normally associated with the word gravitas, I would say.
Okay. So, you watch the the trailer and the first thing that jumped out to me is this is prehaircut Zuckerberg.
Right. The the modern I would say postpandemic Zuckerberg is defined by the haircut the shirts that say like E Plurbus Zuck or whatever they say on them like Caesar era Zuck.
He's gone very Roman. Yeah. Yeah. Caesar era. Yeah. This all the things. This isn't that this is we're going to do congressional hearings. Zuck seems very much like the Facebook files are going to be an important part of this from the trailer. There's one part where you know the person leaking says, 'Are you a tech reporter?' And the person says, 'Is she goes ish?' Which is a perfect sorcin moment. And I'm not sure that ever actually happens in real life.
Also at the end of the trailer where they're like, 'She's a disruptor.' And I was just like, 'I can't.' As somebody who's worked in tech, so like I'm like that word I can't. I just I feel like it's just really I don't know, guys. I don't know if I can do this again.
When you watch this trailer, did you catch that Jeremy Strong is just playing Kendall Roy again? That was 100% my takeaway from this.
Kendall Roy with a really good Mark Zuckerberg accent.
I'm a free speech absolutist. I'm not the one who's lying and I'm not stopping them from seeing someone who is.
He got the voice exactly right.
The voice is really good. Like I am incredibly impressed by that. Uh but yeah, it like especially that moment where he's yelling at Bill Burr. It's just he's like when I say no, that's the end. Like that's the end of the thing. Like you don't get to like that's the end of the conversation. And it's just that it was very Kendall Roy to me.
Yeah. And this is such a weird story to choose. Like this is they're they're essentially making this movie about Francis Hogan and and the Facebook Files. um which is just an odd way of framing this version of the story. Like it'll be interesting to see how uh much it gets away from that. But like Francis Hogan, who like famously moved to Puerto Rico to get away from crypto taxes, is like the the hero of the story. The the whole Facebook Files thing is is very strange and Francis Hogan's story is is strange and complicated. like what came out in the Facebook Files is big and important and like there's almost part of me that wishes Aaron Sorcin had just like straight up made a Cambridge Analytica movie just to like fully piss everybody off.
That's Yeah, that was I was very surprised that this was in fact not a Cambridge Analytica movie. I was like, 'Oh, oh, they're doing Oh, they're doing that scandal, that massive Facebook scandal. My bad.' Like I But what I've been trying to figure out is like do you think Aaron Sorcin just feels bad for like making Mark Zuckerberg look uh cooler than he was at the time and this is why he made this movie. He's like he's like I lionized Facebook by accident and now here we are.
