Hello and welcome to the Vergecast, the flagship podcast of 5G connections. I'm your friend David Pierce, and today on the show we're going to talk about phones, kind of. We're going to talk about this new set of AI devices that appear to be coming for us that are not phones. At least the companies who are making them claim that they're not phones, but they seem to be a lot like phones. They're AI-centric. They are in some cases made by companies who want you to use their AI products, and they're trying desperately to unseat the smartphone while seemingly discovering that the only way to do that is to make a smartphone. David Al, who's with The Verge this summer, is going to come on and we're going to talk through what OpenAI, SpaceX, and Amazon, and potentially many others are up to, and whether any of them can build anything that isn't a smartphone that actually makes you replace your smartphone. All that's coming in just a second, but first, here's what else is happening on The Verge today. This is 90 seconds on The Verge for Monday, July 27th, 2026.
A bunch of big AI companies including Nvidia, Microsoft, SpaceX, IBM, and a whole host of others are creating a new group called the Open Secure AI Alliance. It comes after NVIDIA CEO Jensen Wong shared an open letter last week arguing in favor of open models in general. It became a whole thing over the weekend who signed that letter and who didn't. That particular drama isn't all that interesting, but the US AI industry is clearly concerned with how fast China's AI is improving. And in part, it's improving that fast because its models are so much more open. That means some big questions are left to answer about how we build and share AI going forward.
Meanwhile, it appears Amazon is gearing up to start competing for real with Starlink for satellite internet supremacy. It filed an FCC application over the weekend to deploy 5,15 new satellites starting in 2028 and said it plans to partner with mobile carriers to offer direct-to-satellite service, kind of like Starlink does with T-Mobile right now. Right now, Amazon's Leo service is way behind Starlink. Way behind. But the company bought Global Stars Network earlier this year and has said it intends to offer its own service going forward. There's even been a rumor recently that Amazon might be building a phone again. I wonder if those things are connected.
And finally, The Verge's Antonio Deanto just dropped his review of the Framework Laptop 13 Pro, which is frankly one of the most exciting things to happen to Windows PCs this year. It's modular, it's repairable, it has good battery life and performance and almost everything else you would want. It is also, thanks to the ongoing RAM shortage, $3,100. That's up 800 bucks since it was announced just in April. Ramageddon comes for us all. You can read more about all this at theverge.com. That is 90 seconds on The Verge for Monday, July 27th.
All right, let's talk phones that aren't phones. David Al is here. Hey, buddy.
The Verge's David Al. Waveform's David Al. DavidL.com's David Al.
Just the David Al. He's here.
Right. Good to have you back.
So, I've brought you here to talk about phones, but not actually to talk about phones. I've just become obsessed with this thing that seems to be happening where a bunch of companies are looking for a way to put AI into a device that makes it the most important thing in your life. And basically, if you're not Apple or Google, you can't. You just don't have an option for a bunch of reasons we should probably get into here. But what you now have a bunch of other companies doing, I think most reported, OpenAI, SpaceX, and Amazon are all believed to be reported to be in the process of making something that they would love you to believe is not a phone, but might have an awful lot in common with a phone. And what I want to try to figure out here is whether a) that's possible, b) what it would be, and like at what point is your not-phone just a phone and what's the line you need to avoid crossing, and c) whether this could even possibly work.
Does this all make sense to you? This seems like a crazy moment that we're in in not-phone phone world.
You ask me if it makes sense to me. The idea makes sense to me. The what-would-this-even-be does not really make a lot of sense to me.
Yeah.
Okay. Good. So, let's talk this out a little bit because I think if I'm OpenAI, I assume the thing I want is a ChatGPT-forward device available to you at all times. Is there anything more to it than that? Do you think I'm missing anything here?
It seems like that. I mean, they're going through all the, you know, they make all of these things where they just build the UI on the fly and they just move through intent-based operating. Like the main paradigm shift that I think a lot of these companies are trying to do is that they're trying to move from a tool-based model where you use your phone as a tool, each individual app is a tool, you jump to the tools, to an intent-based model where you can use natural language and your intent becomes the thing and then the apps become the piping for the thing. But if you make a wholly new device that just tries to ignore all of these apps, what are people actually going to be doing? Right? Because we use ChatGPT and we use Anthropic to like create things to then put on the tools that we use every day, right? And this is where I get confused about what exactly are they trying to make, you know?
Yeah. So, I think even as a user, I struggle with the question of what this would look like in a way that I find compelling. And I think you're right that the thing they're trying to do is upend the idea that the first thing I do is open an app to do something, right? Every time I want to do anything, the first thing I have to do is open an app and log into that app. And I think there's actually a pretty strong argument to be made that that sucks. That actually it's really inefficient and remembering where everything is is really complicated and doing things that might exist between multiple platforms is vastly too complicated. I am a long-term believer that Apple Shortcuts is a terrible solution to a very real problem.
Makes some sense to me, but I don't think I just don't think what I want is for my phone to unlock into a chat window, right? Is there something else you can imagine that might look like?
Oh, that's hard. Um, this probably is like the many billion dollar question that a lot of people are R&Ding their way through right now.
Yeah, it probably is. I assume that they probably have a bunch of ideas in a lab somewhere where you open your phone and they try to just guess your intent based on the things that you've done before and the secondary model is to open a chat window. You know what I mean? They probably just want to try to predict what you want to do before you want to do it. But my question again becomes like, well, if apps don't exist, then what do you actually do with a device like this? You know, and maybe apps do exist. You know, OpenAI, their device is apparently going to be an Android phone. Is that sort of the idea? It's like an Android-based device.
That seems to be the running theory, but Android-based device can mean a lot of things, right? You can run a version of Android that has absolutely nothing in common with what most people perceive to be an Android phone. Like your dishwasher is an Android device at this point. So, I think that could mean a lot of things, but I do think there is a case to be made that it might be a regular, normal Android phone with a very ChatGPT-ish launcher.
Yeah. Yeah. I mean, I could kind of see that. The hard part about that is like it feels like the idea of an app and the buttons inside of an app are buttons for a reason, right? Like if you're typing a prompt, that takes like so much longer than just like tapping a bunch of buttons to get to the point where you want to be. If you're on like a multi-chain like do this with this app and then send it to this person and then edit it this way, I could see it being, you know, based around typing or your voice or something like that. But right now, the input mechanism of buttons is actually quite efficient compared to anything else that they might want to use it for.
So it's very true. Yeah. One very simple way I've been trying to think about this is, okay, what if you unlock your phone and the first thing it drops you to is not your home screen of app icons, but the Gemini app. And you can still, you know, you can swipe again and get to your apps. And actually, in theory, they're all there, but we've just sort of flipped the layers, right? That the thing on top is the AI system and underneath it is all of the apps that you maybe want to tap on. Does that do anything for you, you know?
Okay. I'd say that the new Siri is probably a good way to look at this. Now, anytime I want to do basically anything, unless it's literally like texting the first person in my contact list, I just pull down from the top and I ask Siri to do it. Interesting. So, in a way, that is kind of what you're describing. And that means you actually change that behavior fairly quickly. This is the thing I was going to ask you about is how much inertia there is in smartphones. But if you were able to do that just in the course of a couple of developer betas, that strikes me as kind of a meaningful change to make.
Yeah. I mean, I will say that I was at first I was only doing it because I wanted to pressure test Siri as much as I possibly could and see like can it do this? Can it not do this? I don't think it can do this. Could it do it? But it has actually changed my behavior and a lot of the time now I am just asking Siri questions. I mean the biggest question is like Google search revenue is going to disappear because you know I'm just asking Siri all these questions but Google already knows that.
Wait, give me an example. What kinds of things have you discovered you can offload to Siri AI?
Okay, so one big thing, I was on Instagram yesterday morning. So I guess that's not like the first thing I do when I open my phone, but I'm on Instagram and I see that my favorite band is going to announce tickets for the show that they're going to have in December. And it doesn't say anywhere on the screen when they're going to release the tickets. They just announced the show. So I scrolled down. I scrolled down on the Siri thing and I said, 'Can you remind me to buy tickets to this when they go on sale?' So it like searched the web, found when they're going to go on sale, and then it created a reminder to buy the tickets at the exact time that they went on sale, which was this morning.
That might be the single most impressive Siri demo anyone has ever seriously. That is like that's the dream. That's like the stuff they put in advertising videos that never actually work the way they're supposed to. That's awesome.
It's pretty awesome. It's pretty awesome. The other question though becomes like if it's using an Android phone or if it's using Gemini, like Google's kind of like paradigm for these AI-based, you know, for Gemini controlling your phone is that it's sort of a clicking around. So obviously like how the phone is controlled is very important. You know, Apple's version of this is we're going to ask developers to put in all of these schemas and pipes to make sure that this actually works, whereas Google's just tapping around. And so, if you land on like Gemini, you know, you just land in Gemini, is it going to then swipe out of Gemini, enter another app, click on that app? Because the current version of that, as Allison has showed in her article, it like works, but it's very slow.
Yes. And I think it's also going to require a lot of trust. You know, if you just ask the ChatGPT app on your phone to order you Starbucks, there's a lot of trust built in there. Is it my right address? Did it get my order correct? Did it remember to do one pump of vanilla instead of two? I don't know. I think it might take a little while longer. And I'm very curious like what these companies are thinking deep down in their labs on how these interaction paradigms would actually work.
Yeah, I tend to agree. And I think what's so funny about the Siri example is I think you both just described exactly why every company is running at this idea because if you can do enough things like that for people, it starts to feel like an actually genuine paradigm shift in how we use our phones. And also Apple is never, ever, ever, ever going to let anyone other than Siri do that thing, right? And so if you want to do that thing, you have to build something else. And this is to be perfectly honest like I don't spend a lot of my time feeling sympathy for OpenAI, SpaceX, and Amazon. But I do sort of feel for them in this case because you talk yourself into this long enough and you get to the point where to say, 'Oh, to do that for David ML, we have to have an Instagram app.' And then you've built a smartphone. Like you just did it. You do that three times and you have made a smartphone.
And you can see why everyone talks their way into phones. It's like I remember talking to Carl Pay, the CEO of Nothing years ago when they were starting and he was like, 'I have, you know, giant ideas about the future and we want to build lots of new things.' I'm still waiting for any of those to come to fruition. But my question to him was, 'Well, then why make a phone? Like, you think you know what's next? Why make a phone now? Just start going and doing the next thing.' And his answer was just essentially, 'We have to make a phone. If we don't do the phone, we can't do any of the rest of this. We have to make the thing that's in your pocket. We have to figure out how to make all of this work.' And their answer has been build the ecosystem with, you know, earbuds and all this other stuff, then try to sort of back that into some stuff that only works on the Nothing phone and then presumably start to try and flip the form factor. So you like have to do it in a bunch of steps, but you do kind of have to make a phone first, right? So, do you think it's possible that at the end of this OpenAI and Amazon and SpaceX, if they all ship something, it might just be an Android phone or something essentially identical to an Android phone?
I think it's essentially going to be an Android phone. I mean, we saw the Rabbit R1 like try to do something. Now, given that was super early and it didn't work and what they said it did, it couldn't really do, but that was also just an Android phone. You know what I mean? It was like we made it a different shape. And you know, it's like are there models embedded enough into your life for your daily interaction paradigms that you are going to want to use that model to interact with those things that you do every single day, right? I don't know. I mean, Google has the best shot with Gemini. Apple has a very good shot. You talked about the earbuds with the cameras. I don't know what their interaction paradigm with those cameras is going to be, but they already like if Carl P said, 'We have to make the phone first and then we can make the smart things that attach to the Nothing phone later.' That's exactly what Apple's doing.
A third party coming in and doing that, it's just like, is anybody that invested in Grok that they want a Grok device? Like I don't, I just, you know, I don't like being pessimistic and being like there is a duopoly, but there is a duopoly.
Support for the show comes from Even Realities. For a long time, the reality of smart glasses was that they were these bulky, inelegant pieces of tech that were much cooler in concept than in practice. A far cry from the sleek, stylish accessory of our sci-fi dreams that actually provides real functionality. Well, Even Realities has that fixed. Even G2 are productivity smart glasses designed to keep real-time support right in view with teleprompting, conversation support, real-time translation, AI assistance, and more. They help you stay on top of work and daily life. And unlike most smart glasses, they're designed to look and feel like premium eyewear with no camera and a lightweight 36 g design you can wear all day. The more context you give them, the smarter they get, adapting to how you work and what you need. To learn more about Even G2, go to evenrealities.com and see how everyday smart glasses keep helpful information in sight so you can stay productive and hands-free throughout the day. And for our listeners, use promo code verge at evenrealities.com to get 10% off Even Ring 1 and/or Even Clip when you add them to your Even G2 order. That's evenrealities.com promo code verge.
Support for the show comes from MongoDB. AI-assisted and agentic coding can help you build faster than ever. But if your data layer is a bottleneck, what's the point? MongoDB actually gets out of your way. MongoDB is a unified AI-ready data platform that empowers you to build scalable generative AI applications. It eliminates the need for separate specialized vector databases by combining a flexible document model natively with semantic vector search, full-text search, and real-time operational data. Instead of wrestling with rigid schemas or translating data formats, MongoDB's native data model mirrors the language LLMs already speak. Plus, MongoDB gives you the flexibility to ship at the speed of AI. The ACID compliance lets you sleep soundly at night and scales to handle massive Fortune 500 workloads. Developers have a word for that kind of reliability. Actually, five words. It's a great freaking database. Start building at mongodb.com/ai.
At this point, particularly in the US, there are only two phones that matter. There's the iPhone and there's whatever Samsung shipped. That's it. The people like I love the Pixel. I think more people should buy the Pixel. People don't buy the Pixel. People buy Samsung phones and they buy iPhones.
Is it even conceivable to you that OpenAI could show up with a pretty good Android phone and make a meaningful number of people switch?
What was the positivity rate around AI in the United States again? Like lower than the Democratic Party?
I don't think they can do that. They might get like a quarter of San Francisco maybe.
Yeah, but they're not going to get anybody else.
And also, you know, the problem that that puts you into as well is that if you buy a device made by a model maker, it's inherently if not only going to work with that model maker or works best with that model maker. And it doesn't really feel like a lot of people have allegiance to any particular model. I guess there's a little bit of allegiance to like Claude, but that's because they're supposedly the company that's trying to do the right thing.
Well, ironically, I think the allegiance to Claude is as much a referendum on people don't like OpenAI as it is anything else.
That is also very true. That is also very true. But either I mean, as soon as I, I don't know. I'm a big believer in like these open source models are starting to get crazy and as soon as RAM gets cheaper again, I think people are going to start running their own, especially at business scale. You know, if you can buy RAM cheaply again, a lot of these open source models are almost as good and you just really don't need to use any of these model providers. So, I don't know. A device for me makes no sense. I think a smartphone running what these companies want you to use makes sense and it is the most basic thing that we've always had, but disrupting that interaction paradigm is like I don't see how you do it. By the way, like you were saying, you either open up to a chat box, which is kind of annoying to open up to every single time you unlock your phone or you have to talk to it. And the reason that things that you have to talk to in order to interact with them don't work is because we live in a society. So, you know, the Vision Pro very isolating. Everything's isolating. Isolating things don't work.
I will say I am extremely bullish on doing more things on your phone with voice and fewer with a keyboard. Good dictation on a smartphone is life-changing. Like I did this big experiment earlier this year where I tried desperately to get rid of my iPhone. Tried all the other Android phones and then came back to the iPhone because I'm a deeply broken person.
But you and everybody else.
The biggest thing I discovered, my absolute favorite thing about the day-to-day experience of Android is how much better dictation was.
Oh my god. Gemini is better. Voice recognition in general is vastly better on Android. But my ability to successfully type text messages on an Android phone leaps and bounds beyond what I've ever experienced on an iPhone. And it was wonderful.
And I think in general, I've been testing all these dictation apps on the Mac too. And the idea that they're starting to understand which app you're in and actually communicate and transcribe differently based on the apps, I think that's pure AI and that's very powerful and is going to be very exciting. That is OpenAI showing up and being like we have a better dictation model on our phone, you should buy our phone instead is not going to work.
But Amazon is the one of these that I'm the most stuck on because right Amazon 2014 tried the Fire Phone. Obviously didn't work and it was full of bad ideas. I think the most bad idea being that what people mostly want is a closer connection to Amazon.com all the time. That is just not correct. And then there's been this reporting that Amazon is building a phone again. It's internally known as Transformer, which is a rad name for a phone. They should just call it that. But let me just read you two quotes. This is from a Reuters story about this phone.
It says, 'The phone is seen as a potential mobile personalization device that can sync with home voice assistant Alexa and serve as a conduit to Amazon customers throughout the day.' The people said, quote number two, 'As envisioned, the new phone's personalization features would make buying from Amazon.com, watching Prime Video, listening to Prime Music, or ordering food from partners like GrubHub easier than ever.'
Amazon has learned nothing, David. They're just doing it again.
Does anyone really want Capitalism OS? That's basically what that is. It's just like, oh my god, I don't watch TV really at all. But I was over at a friend's place and they had a Prime TV show on. I think they were watching Fallout or something. I did not know that there are literal buttons on the ads that you can just buy and add to your Amazon cart while you're watching the TV show.
