The Vergecast - AI can't even turn on the lights

Episode Date: October 17, 2025

Nilay’s back! And you can listen to The Vergecast with no ads, if you’re a Verge subscriber! Big week, really. Nilay and David start the show by talking about ads, podcasts, platforms, and subscriptions. Then they talk a bunch about Apple’s new M5-powered MacBook, iPad, and Vision Pro, and whether a chip bump is worth getting excited about. After that, Nilay reflects on a summer of using AI products, and explains why you can tell the whole story of this generation of AI just by talking about the smart home. Finally, in the lightning round, the hosts talk about AI song covers, Apple TV, TiVo, Roku, Cybertrucks, and the exploding Pixel 10 Pro Fold. Help us improve The Verge: Take our quick survey at theverge.com/survey. Further reading: Ad-free Verge podcasts have arrived Netflix is making a big bet on video podcasts Apple’s 2025 iPad Pro comes with an M5 chip inside  Apple just upgraded the Vision Pro with an M5 chip and new strap  Apple’s 14-inch MacBook Pro gets an M5 chip bump and faster storage  Logitech made an Apple Pencil-like stylus for the Vision Pro  Apple’s rumored smart home display hub might start at $350  Samsung officially teases Moohan headset launch for next week  Apple’s future smart glasses could have two separate UIs.  ChatGPT will soon help you shop at Walmart.  How OpenAI plans to make all its money.  Microsoft wants you to talk to your PC and let AI control it   As Microsoft bids farewell to Windows 10, millions of users won’t  Spotify says it’s working with labels on ‘responsible’ AI music tools  DirecTV will soon bring AI ads to your screensaver  OpenAI partners with Broadcom to produce its own AI chips  Sam Altman says ChatGPT will soon sext with verified adults  Apple TV Plus is being rebranded to… Apple TV  Apple exec on Apple TV rebranding: ‘let’s just do it’ Google’s Pixel 10 Pro Fold is the first to ‘go up in smoke during a bend test,’ JerryRigEverything says Roku’s AI-upgraded voice assistant can answer questions about what you’re watching  DirecTV will soon bring AI ads to your screensaver  Soul Against the Machine TiVo has sold its last DVR  Tesla Cybertruck sales are flatlining  Subscribe to The Verge for unlimited access to theverge.com, subscriber-exclusive newsletters, and our ad-free podcast feed.We love hearing from you! Email your questions and thoughts to [e-mail] or call us at 866-VERGE11. Learn more about your ad choices. Visit podcastchoices.com/adchoices

Support for the show comes from Retool. Too many companies run critical operations on duct taped spreadsheets, Slack workflows, and whatever else they could cobble together. Not because they want to, but because building internal tools means weeks of waiting on someone else's backlog. That's where Retool comes in. Build custom internal tools just by describing what you need. Prompts something like,

Build Me a Revenue Dashboard on our Salesforce data. And Retool actually builds it on your company's data, in your cloud with enterprise security built in. Go to retool.com slash Verchcast. We all need to retool how we build software. What's up, y'all. I'm Skyler Diggins, seven-time WMBA All-Star, Olympic gold medalist, and mom. And I'm Cassidy Hubbard, host and reporter for nearly 20 years,

covering the biggest names and stories in sports and mom. And this is Am Mom, a community for athletes, game changers, and moms of all kinds. dropping May 14th. Tap in with us. Do you ever wonder what's in your lotion? If you look at the back of the bottle, it could contain more than a dozen ingredients.

And they may not all be regulated. The threshold is so high that only 11 cosmetic ingredients have been restricted by the FDA since 1938. This week on Explain It to Me, the chemicals lurking in your cosmetics. New episodes, Sundays, wherever you get your podcasts. Welcome to the Vergecast, the flagship podcast of Internet-connected baby monitors. I'm your friend David Pierce, and he's back, my friends, the one, the only, the man, the myth, the legend.

Neil I Patel is here. Hey, buddy, welcome back. I didn't prepare. Which is very different from other episodes. Everyone should know. You normally come just prepared to the teeth this time. Just full of riffs. No, I'm back.

I'm back. It's week one. you know. How is it being back? It's been a very weird transition for me. Yeah, I'm curious how it was for you mean, you're like the skeleton key, right? I was like, how's baby too? And then you gave me all the tips, which were it's fine, basically.

More or less. Yeah, it's been like, it's been just pure, unadulterated chaos, but I'm like used to it now, if that makes sense. Like, I know how to do the chaos now. And so it's been, it's been okay. But then they'll be teenagers and they'll change. I was going to take way more time off. And then our baby, Jack, just started sleeping through the night.

So my job was to do all the nights. Because what I learned from our first baby, Max, seven years ago, was that if Becky doesn't get enough sleep, we are all in danger. Physical, emotional danger. Yeah, it's not good. And then like a seven-year-old versus sleepless wife is like a, the town is in danger. Like, that has a blast radius. So I was like, look, I'll just do it.

I like being up at night. I can scroll TikTok. It's fine. I'll just be up all night with this baby. And then we'll do it. And then literally, he was like, well, I sleep now. So then I was just buying stuff on Instagram by myself in the middle of the night. And I was like, I should get back to work.

Yeah, nothing to do. What's the dumbest thing you bought on the internet during this process? We own a steam mop, David. Why do we own a steam mop? It's a mop that steams. Does it mop with steam or does it mop and also steam? It's unclear if it accomplishes either of those goals, particularly. Well, just why?

Why? You know, it's like you watch a cleaning video and you're like, I'm going to get that bathroom real clean. And then you are in your bathroom holding a steam mop. And it turns out they produce steam. So you're like, I'm hot and damp alone and I've made a mistake. I should go back to work. Like, I had that full, you know, they don't tell you that it's damp in the Instagram video. They're like, you will be more damp than you were before.

for if you use this product. Yeah, Steam, who would have thunk? You know what I mean? Who could have possibly seen that coming? Also, just, you know, I don't know who you're this probably like, I'm a sucker for a cute baby onesie. Oh, God, yeah. Lots of those.

A truly alarming number of, like, brand name onesies have come across. I own a Nike swoosh Just Do It, newborn buddy now that I'm very excited about. Has you grown into or out of it yet? I mean, one and then almost immediately the other. Yeah, that's how it goes. just booked it through it. So, Vergecast's question for you, because I've actually been asked this a bunch.

Did you do any tech stuff differently this time than last time? Actually, you and I were texting about this with our friend Joanna. So I thought we would upgrade the baby monitors, and I'm totally opposed to having internet-connected cameras

in our house. Just not a thing that I want to screw with. So I was like, I'll get ones that do like the picture and picture or the split screen, and we screwed with it and that was too much user interface you know?

Like our daughter, she doesn't want that baby monitor room but she wants to yell us. She uses it like a walkie-talkie. Sure. Right? Like that's what it's there for. She's in her room and she's like, dad, and it comes through on the thing.

So instead we just have two. And one's labeled Max and one is labeled Jack and the hardware walks around. And literally my wife and I refer to them as the children. And I'm like, do you have Max? And she like waves a baby on her me. So that was like the big, I thought we would solve the problem with, like, innovative technology,

and we just solved it with adding more devices. We just threw hardware at the problem until it was solved. But no weird sensors or anything like that? Like, I feel like the new thing that happened since Max was there's now just a billion sensors you can put everywhere. Yes. So, Jack was in the NICU for the first nine days, and so I bought an outlet, and then I immediately returned the outlet. Like instantly.

For like a whole number of reasons. First, I think I was just annoyed that it only runs on 2.4 gigahertz Wi-Fi. Just in principle, yeah. That has to get out of it. Right, like being like to set up this sock that will do nothing but false alarm, I have to put my ERO, like the whole network has to be downgraded to 2.4 gigahertz for 90 minutes so that it can figure itself out. And I was like, I hate you. And then, you know, it's like they're not even meant for very small babies.

which is not a thing anyone says out loud. Yeah, I did not know that. Here's my outlet review. They're not meant for very small babies and they will constantly false alarm, which is like when you want it the most, really. So whatever. So we just like immediately get that back.

The sensors that we did add is the smart home is out of control in this house all this time. Okay. So he's got automatic shades in his room. And I have a lot to say about shades. So when you close the shades in his room, the lights go out.

off and the fan turns on. Hmm. And then when you open the shades, the noise machine turns off. Oh, okay. So there's just like a lot of if-then statements in his room. It's like the good night and good morning routine, basically. Yeah, because otherwise the noise machine is on 24-7

because no one can be bothered to turn off. You're holding the baby and it's drive me nuts. So that's on a smart plug. All the other stuff is on Lutron's, which is, and it's all just like, I'm sitting here just like writing if-then statements in the rocker while the baby's sleeping. like, okay, if this happens, what four other things should happen, but conditionally, this other stuff should happen.

A bunch of matter stuff made its way, like matter and thread devices made its way into our house. Huge win for Gen-toe. Somewhere Jen is just fist-pumping, listening to this. Huge way for Jen-toe. Not a win for those companies, because now I know how well those devices work. And it all feels like a hack. So the way it works, more smart shades. The shades in his room are matter over-thread smart shades.

They're Canistaya ones that I bought on Amazon on the recommendation of Stephen Robles, who runs a great YouTube channel. And we'll be on this podcast very soon. He and I are scheduled to fight about shortcuts. Yeah, it was great. I was like, I'm buying these, and he's like, he's made videos

about every single one you can buy. They are all rebranded smart wings. Like, at the end of the day, they're all the same. But these are great. They're inexpensive, hung up. So to do it, you've got to pair them to HomeKit, and then HomeKit will generate a code that you can have Google Home pair to.

So it's like these weird round trip where Apple becomes the hub for Google and everything. And it is like, oh, this is not great, actually. Like the thing that you think is going to happen, which is you have multiple smart home platforms all working together in harmony, is much more like I'm typing strings of alpha numeric characters into Google and praying that this works until I realize that I have multiple competing thread networks in my home. And I once again have to shut my ear over. down to like it's a base configuration to bring it back up again. But it all works now. It does all work now.

Until like the light changes and it's all over. Like a gentle breeze will screw up your smart home at this point. It's all rock solid and Apple Home. And it is all Google Home is like, what? Who am I? We'll come to this. This is my theory about Google and AI. But like the Google Home stuff is like the canary and the coal mine for the entire bubble to pop.

Because it's just every day it wakes up. And it's like, I don't know who you are, and I don't know where I am, but I will tell you facts about the moon. You know, it's like, what's going on? What are you? Did you do any stuff? Not really. We did sort of the same baby monitor thing you did.

So we got a video monitor, but I think it's VTech. I think it's the company that makes it. That's not connected to the internet. It's just, it's all like on our local network. That's been great. We just got another one of those. And now it's split screens and that works great. We were going to do a bunch more stuff, but then immediately started house hunting.

So then it was like, okay, well, we're going to wait and do all our smart home stuff. We're going to have to, you and Jen and I are going to have to have like a long smart home therapy session as I move into this house because I have lots of questions. And also a spouse who is not interested in any of this stuff. So I have to do it all in like a sneaky way that she doesn't even notice. That's the best way. We'll see how that goes. But we ended up doing less than I expected. Like with our first kid, Arthur, there was a period of time.

when he was like two weeks old, where I basically went on Amazon and bought every sleep device that existed. And I was just like, whatever amount of money it costs me to have this child sleep and whatever weird Bluetooth technology it requires, I will do it.

And our Lewis, our youngest, he just like sort of did okay from the beginning. And so the like panic, I have to buy a technology to solve this problem, problem just didn't really appear, which was nice. We got lucky on that. The flip side has been

I am much more susceptible to being conned by Instagram and TikTok ads late at night than I used to be. Do you have a steam mop? I don't have a steam mop. I got really into a Swedish candy over the summer because somehow that ended up on my timeline a lot.

It's just like this is just like, sure, that sounds delicious. You can ship that to me in three days, Bon Bon. Like, let's do this. Yeah, it was just it was a lot of scrolling in one hand. I have become even more of a pop Sockets, zealot, because you can hold it in one hand while you hold a baby. No notes.

Oh, this is a game changer. I have dropped my iPhone 17 Pro Max every single day. This is what I'm taking. So there's the new pop socket with the kickstand now that you can actually, not only is is their pop socket, it folds out so you can prop up your phone on things. Do you know how many things I prop up my phone on, Eli? All of them. All I got for you is that it's very good that you're married because if you were not, the chances

procreation would drop to zero. No one ever suggested that I was cool. But my phone never falls down anymore. I don't know what that thing is, man. My job in my marriage is to be the one who brings all the chargers, and I do that very well. Yeah, that's correct. That's about where we are.

Anna just does not know cables exist, and she just, whenever her phone dies, she just reaches out her hand, and it's my job to have the correct cable. That's great. And sometimes it's a USBC cable. Yeah. I mean, I just like have a bag full of them.

Yeah, same. That's great. All right, there's a bunch of news we should get into. There's some Apple stuff to talk about, ish. There's a bunch of AI news, as always, we've got a lightning around to do. We should probably do just like a little bit of housekeeping up the top here. Most people listening to this or watching this probably already know. But we just launched ad-free Verge podcasts.

If you're a Verge subscriber, you can now get all of our podcasts. this one, Decoder and Version History, all ad-free. There's a post on the site. We'll put in the show notes with the instructions on how to do that. But, Nil, I want you to talk about this because we talked about this when subscriptions were coming out. And basically since, like, day one of The Verge has a subscription, the number one thing people have asked us for as ad-free podcasts. And I got a post, I think, on threads from somebody this week who was like, well, what changed? You guys made a very compelling case for why this is difficult to do from a business perspective.

doesn't really make sense. What's changed? So, like, why did we do this? What's your take? A few things. One, we're coming up on a year of the subscription. So subscription businesses, I think this is obvious to anybody who's paying attention to it, certainly for Chast listeners. Subscription businesses are all about renewals, right? You're obviously trying to get new people to subscribe, but if you turn off a bunch of people that you already got, then you're in trouble. So we're coming up in a year. We wanted to, we have a lot of ideas for how to make even better for all the people to subscribe. And this was, like you said, by far, the number one request. I'm happy to pay you. I don't want ads in this podcast. So you just look at the list.

And, you know, if you were making a bar graph, it would be like the one bar is the tallest. And that ruins the scale. And like, everything else is nothing. Like, that's basically how this, the request broke down. So we figured we had to do it. There are new tools to do it that sort of like don't ruin our analytics, which was a big. thing. Oh, interesting. So that's interesting. We'll see how those play out. And then we were able to figure out basically how to make the money work inside of the subscription. This is like how in our publisher's zone. But I think our overall plan is to grow. Right? We want to make the podcast more popular. And so if you have subscribers and are helping you grow and that evens out the ad market,

which is the first thing to go if you are worried about the economy. Yeah. Then saying, okay, we're going to bet on subscribers and try to grow the overall pie to be more lucrative to the advertisers. You do show up. Like, you're at least stable, right? You diversified the revenue. And I, my perspective is like, we need to do that. We should do that a lot. I'm looking broadly at the ad ecosystem, the internet ecosystem, and everyone is turning

into a little ad agency. I think you guys were talking about unwell marketing and pixel phones. Yeah, it's Cooper's thing. Alex Cooper's thing on the last one. Alex Cooper showed up at Cannes, which is like the big ad festival, and she said out loud, I'm not an influencer,

I'm a marketer. Great. Again, everyone should go get their money. I have no complaint. I'm not going to fault anybody for making their bag by playing the game. It's kind of designed to be played.

Yeah. But she's a marketer. That is what all of those platforms want you to be. Like, to be super clear, that is not Alex Cooper like making outrageous claims about how things work.

That is Alex Cooper saying everything about this system demands that I be a marketer. I am a marketer. Right. Instagram wants you to shop. Yeah. We started the show by being like, we shop on Instagram. We have a bunch of little QVCs. We, you know, we don't want to do that. Like, I don't know how I was to say it. Like, I want to make journalism. I don't want people to be able to pay us to say things or make ads. I don't want to make, you don't want me making ads. No. We did that one little run where

they, like, had me read the ads and the advertisers were like, make him sound happy. And I was like, refused. Richard Lawler on our team the other day just randomly there was some like LeBron James Hennessy thing

that happened and Richard reminded me of the time that I read a Hennessy ad on the Vergecast and told me that that brought him

great joy for a very long time. There was one, there was a period where I told our ad sales team that the only exception was liquor ads.

I would read as many ads for views as anything they wanted and that is the one that they delivered. I don't know if that was a good idea or not, but we had the experience, basically, and the idea that we would do branded segments

or we would even just read the ads. Like, I think, I think when you are buying our subscription, what you are buying is our ethics policy. I mean that very sincerely. That is the thing you are buying from us. Yes, we make fun podcasts. Yes, the site is full of interesting stories. Yes, Liz is occasionally prone to just tearing someone to bit, like limb from limb into tiny

little shreds. and that is a worthwhile thing to pay for. But the fundamental thing that I think we are selling is editorial independence. Yep. Which is what enables us to do all of that stuff, right? Like, it's not like some high and mighty thing. Like it is the thing that enables all of that other stuff.

Yeah. And that, to me, you know, on my break, staring at my phone, consuming a lot of platform content because scrolling with phone in your hand, no pop socket, I will say. But scrolling, holding a baby with phone a hand, yeah, I consumed a lot of platform algorithmic media. And I was like, all this is ads. at the end of the day

everyone who is great at this is doing brand deals and if you are prone to perceiving it you're like I can see I can see all the pressure that is being placed upon all of these folks and the way that I see it particularly in tech is the dam doesn't break until it breaks

on like conventional wisdom because everyone is like no one wants to be the person who does the criticism and doesn't get the brand deal down the line and then the damn breaks and everyone's like, this is bad. And then everyone is like, this is bad because that's all the views. Right. And I have millions of examples of this. But like, there's just a piece of that puzzle that I think is, I don't want to play the game. Anyway, to bring it back to Ad Free Podcasts, I think we're able to launch it because we, we, the subscription is working and we want to make it really valuable for people to renew, really valuable people send up.

The overwhelming feedback from all of you was give us ad free podcast. And then I think it's very important to not only just diversify the revenue because you think the economy is going to do whatever the economy is going to do, but to make sure that we preserve our editorial independence. To say the subscription is so valuable that the pressures of the ever-changing ad market on platforms will forever be distant to us. So we'll do some other stuff. I have lots of other ideas for like how to make that case on these platforms. I think it will probably get us in some trouble. But the thing that you're buying is our editorial independence. Well, we should, you know, so many people have a relationship with you and me and the rest of our team through these podcasts.

The podcast should reflect that too, I think. Yeah, I agree. Yeah, and I think A, thank you to everybody who's reached out being excited about this. Like, it's very, it's very rare that you just like announce a thing and everybody is just like, oh, thank God you finally did this thing. It was nice to be able to do that. I don't know what your replies and DMs have been like, but just like, universe. universal joy, that this is like a thing that exists has made me really happy. I do feel like we pulled the Apple move where we like left off one feature.

We made the keyboard work again. Yeah. And then like we're like a year later, like we had this great idea. What if the camera's zoomed? Do you remember USB ports? Do you want one of those? It's like, oh, this is why they do this. Everyone's going to congratulate them because the phone is finally waterproof.

Like, okay, fine. Great. It's pretty good. Okay. So speaking of Apple, let's get into some of the news of the week. Yeah. The big product news of the week was on Wednesday, Apple announced a bunch of new stuff, all unified essentially by the new M5 chip, which the fact of that is fascinating.

And we should talk about it. But basically, new iPad Pro with an M5, new 14-inch MacBook Pro with an M5, new Vision Pro with an M5, just like pretty simple sort of internal revs of all of these devices with a new chip inside. Any of these get you excited? Not really. I even saw one of the Apple sets. I can't remember which one at the top of my head. Their take was the M5 MacBook Pro is such a small incremental update.

It is the best case for waiting for the next generation. So if you're Apple, then why ship these things at all? Like, this is kind of what I was running. The iPad Pro at the M4 is still the best tablet anyone will sell you anywhere. The MacBook Pro doesn't have any particular competition. The Vision Pro, you know, LOL. Like, why? I did sort of.

wonder why these things came out in this form at all? My read, a very big remove, right? I've been out of it for several months. I didn't go to the briefings. I think we have reviews coming. I don't know if Apple will address any of this directly. But my read is that they are managing their supply chain very tightly. And they are moving manufacturing around.

And in particular, if you have this big bet on TSM and all of your ships are made in Taiwan, And there's all kinds of tariff stuff happening. And you just get the feeling that they needed to move some chip capacity from one place to another. And you can't go backwards. Right. So you might as well move, you might as well start the M5 stuff and get that going. And maybe you can reallocate some of your M4 to your more popular products. Because I don't think this base MacBook Pro is one of their most popular products.

I doubt it. Yeah. The Vision Pro is not one of their most popular products. The iPad pros may be a different case, but I think it's so expensive. I don't think it is among the most popular iPads. But in its zone, I think it is more popular than the other two. But you just get the sense that they're kind of like,

okay, we can make these updates. We need to start this production, but there's so much other stuff we have to worry about that we should make sure we have supply for that. Yeah. That is my sense. Like a total outside, this is a shot in the dark guess, but like why update these three products

with this trip, well, I just kind of like look at the world. And I'm like, there's a, they are managing that supply chain against tariffs and trade and China trade. Like, this is the kind of thing you start doing where you just start updating products off cadence to create capacity elsewhere. Yeah. Well, I think that is certainly true. And it makes very clear that Apple is now this company that we've thought of for so many years as run by this, like team of brilliant designers standing around those blonde wood desks. And it's like this very sort of iconic image that lots of people have about how Apple works. This company is run by the supply chain people and Johnny Shrugi and his team in that

like weird lab making Apple Silicon. Like the cadence of this company are based on those two teams. Yeah. At this moment. And so it's like what seems to have happened. And this is what we see from lots of other like companies that make sort of constant iterative technology is they base their product revs on new chips, right? And so like when you get a new Dell, it's not because Dell had some wild new idea about what a laptop should be. It's because

there's a new chip to put in it. And so they, they upgrade when there's a new chip to put inside. And increasingly with a lot of these products, it seems like Apple is willing to upgrade when there is a new chip to put inside and then to make big bets on what that chip can do. And like, with a lot of this, I mean, you even read Apple's press releases and they're like, they're basically like, well, if you had an M4, it's fine. You're good. Don't worry about it. But then they talk a lot about like the M1, right?

And it's like, if you bought one of these devices four or five years ago, this is going to be a huge upgrade. But the huge upgrade is like it'll do local AI inference tasks two and a half times faster. Wait, can I just say that's more specific than what Apple is saying. Apple says the M5 in the new iPad Pro brings up to 3.5x the AI performance. of last year's pro. I don't know if you were called the AI performance of last year's pro, but it, it was nothing. Yeah. Like, it's just like, nothing happened there.

They summarize some notifications badly. Yeah. So, I mean, even that is like, are we running local models on this thing on your iPad? Well, you're certainly not doing in the background. Like, what are we talking about? Yeah. Like, I think, honestly, my other read on this is the M5 must be cheaper to produce. And the new networking chips, the C1X cellular modem, if you get that in the iPad, the N1 chip, their in-house networking chip for Wi-Fi, Bluetooth, and thread for some reason.

All that must be cheaper than the ones they were buying because they're no longer paying, you know, Broadcom or Intel's or Qualcomm's fees. So, like, maybe they're just moving to the new architecture, which is cheap. Yeah, and they're like, if we put this in new devices, we can get it at scale, we can make it even cheaper. The race to make that across the line makes a lot of sense. But it is, it's just like what an incredibly rational and boring set of reasons to roll out new devices. Well, can we just talk about the Vision Pro? I feel like the MacBook and the iPad getting chip bumps is like, whatever. Like, you're right.

Dell does it. HP does it. Fine. It's a laptop got a chip bump. There's almost nothing to say. The Vision Pro, it got a chip bump. Great. It went from the M2DM5.

It got heavier. Did it really? It got 150 grams heavier. And they changed the strap. And I think the strap, it's not like a heavier strap, but it's the dual knit now. That's the one they're doing, which I should have shipped from the beginning. Yes. Because it is much more balanced.

But they made it heavier. And they're adding basketball to it. yeah there's going to be some like immersive NBA games starting next year which a bunch of people I've seen are excited about and why like this is a thing everybody who does this tries like meta did this years ago with the quest where they were like you can sit at half court and watch an NBA game and it's like well first of all on the floor at half court is actually like a pretty bad seat so you're going to give me a bad view with none of the fun things about sitting court side it's can

to be like blocky and blurry. I'm going to have to go like this back and forth the whole time while I'm watching. This is not actually a good viewing experience for what we're doing here. It just doesn't, it's like not a, it's not a thing that people want. Like we're pretty good at TV sports now. I don't know. I don't know that the next gen of TV sports is three, you're, you're in the sports. I don't know. And so all of these things, like Apple keeps rolling these things out sort of one by one where they're like, you know, they made an immersive documentary or they'll do a music video or whatever. And every single one of them,

the overwhelming reaction is like, oh, neat. And then no one ever talks about it again. Yeah. And none of them are sticky, right? I mean, maybe you can watch a season of Esquil and immersive 3D or whatever will be very, very sticky. Every time they show the demo of whatever F1 app where, like, the track is on the table and the cars are going around the table.

I'm like, that would be great if I had a huge house with empty tables all over the place. Right. And eight friends who all had Vision pros in the room with me. But the chip bump for the Vision Pro, animated heavier, I feel like it's just the wrong direction. What I want is a cheaper, less powerful, lighter Vision Pro.

Yes. And this thing is just going the wrong direction. I just Googled how heavy is 150 grams, just because, you know, it's helpful to have a vision. I found a website called measuringly.com. Sure. Fabulous that this exists. Here's some things that weigh 150 grams.

would you like to hear them? Yes. A cup of all-purpose flour. A cup of almond nuts. One and a half cups of oatmeal. Four slices of bread. Three strawberry pop tarts. All right.

So the next time I see you wearing a Vision Pro, I'm just going to put four slices of bread on your head and be like, is this better? Another good one is an iPhone SE. That's great. It's one iPhone SE. Yeah. Three pop tarts, one iPhone SE and four slices of bread.

Sure. I mean, like, it's fine. Like, none of that feels like a lot, but it's on your head. Right. And I don't know if the new strap will, like, counteract it or make it better or weight it better or balance it out better. It doesn't seem like they've made big changes to how it balances, right? The strap might distribute the weight slightly differently, but they haven't actually counterbalance it. So it's still very front-heavy. I don't know. It's truly their most confusing product. It is they're updating it against, right?

Meta display is going to start shipping soon. Samsung is teasing the Mujan headset next week, which is their partnership with Google. It's just like this thing, my review is basically like, this is the end of this road. Yeah. Pass through. Big VR pass-through headsets have come to the end of the line.

This is the best one you can possibly make. It's $3,500. And I still feel, I think, exactly the same way. Yeah, my bet for what's about to happen is the meta displays are going to be like a medium-sized hit. Meta's not making that many of them, but I think they'll do well. People will buy them. I think the new generation of regular meta, Rayben smart glasses are going to do very well. I think Mujan is a stupid name, but that thing is going to be almost as good as the Vision Pro and like drastically cheaper.

Like half the price would be my guess. I wouldn't be surprised if this thing is like $1,500 and all of a sudden it's going to make the Vision Pro look just completely on an island by itself. And all the reporting, by the way, suggests that Apple is keenly aware of this and is essentially running away from the Vision Pro

as fast as it possibly can. And yet they're doing a chip bump. I know. That's the odd thing. Like, it would almost make more sense if Apple would do the thing that it occasionally does and just sort of studiously forget that a product exists

where they're just like, they just refuse to acknowledge that it's still there and part of the company, and then someday they'll kill it. But Apple is clearly onto smart glasses as a thing, has pivoted pretty hard internally towards making that stuff work. They've had some weird employee turnover that might make that harder. Meta just keeps throwing giant sums of money at everybody who works anywhere doing anything with AI. But I think we're headed towards a moment here where the Vision Pro just doesn't make any sense to anybody. Yeah, we're already at that moment.

We probably are. You know, my read on the meta display glasses is that meta is equally confused as Apple, but their product is closer to what people want. I think the only reason people are buying the meta ray bands is because they can't run them. And headphones. It may be sometimes the headphones. It's most, it's like, it's like 60% camera, 30% headphones, 10% everything else. If you bought these, let us know.

Send us to note. Right? I'm very curious, but everyone I know who has them is like, I bought these for the camera. No one is sitting around talking to Meta's AI assistant. Meta has confused itself into fully believing that they have found the form factor by which people will engage with their AI assistant. And it's like, no, they just have a camera. But you just made a camera in literally a pair of raybands that look pretty good. And so people bought the camera and the pair of raybands that look pretty good. and then you get all the way to, okay,

if people have the AI assistant, now I have to put a screen on there because everyone wants to build AR, and they're like, now the glasses look bad. They look bad. Yeah, they do. V on our team has been using them and is, I think, getting a review ready.

And she, like, shows up to meetings, wearing them every time, and you, like, look at them and you don't really clock them as anything. And then, like, you sort of, you always kind of double take at them because it's like, wait, why are your glasses enormous?

Did you buy the wrong? Like are those costume ray bands? Like what are we doing here? And it's just, it's sort of like the Vision Pro in a very different way and that it is like, it is very technically impressive that this thing exists.

I don't find it compelling. And from Mehta's perspective, like I think the best argument for what you just said is that as far as I can tell, there's not that much interesting stuff to do with the screen at all right now. No.

Their big idea is like walking directions. Like it'll give you a heads-up display to tell you to turn right. And it's like, I know there are so many devices with so many ideas about how to tell me to turn right and my phone does it just fine. Like, that's, oh, I'm good. It'll be all right.

Yeah. But again, I think they've concluded for better or worse that they have found a piece of hardware that can replace your phone and the interface for that will not be the neural band, which is fine. It's impressive because it is the thing that it is.

But you're like, in order to use my computer, I have to put on glasses in a bracelet. Just, right? Like, my phone is much better at these things. Yeah. And they've decided that this is how they will displace the phone, that the user interface of the future is somewhere embedded in this hardware. And I think they've missed the fact that everyone's just buying the cool ones for the camera.

And I don't think anyone's going to make the leap to, I'm going to start talking to my glasses. Well, and it makes you wonder if maybe Apple sees that, right? Like if Apple called it wrong and tried to do like the Tesla Roadster strategy where you ship the really expensive luxury thing first and then scale it down rather than shipping the minimum viable product version of it and building up. And if somebody other than meta is looking at it and going like, oh, this actually isn't an AI product. It's something else that we can eventually do AI to. There are a lot of interesting ways you could take the thing over time. Sure, but the dream is augmented reality. None of them are actually augmenting anything.

We bought a new car. We bought a catalogistic and has an augmented reality heads-up display. Sure. And really, it has an arrow that gets bigger in your line of sight as you get closer to the turn. This is what I'm saying. Everybody has ideas about turning right. Like, it starts out really far away. And then it's like the arrow zooms in on you.

And it's like an ever larger cartoon arrow is getting closer to you. And you're like, you haven't augmented anything. Like, this is, it's very clever, right? You've tricked everyone into believing that you know. But like, really, the arrow is just getting bigger. Like, very big at the end. It's very funny.

But the goal is, like, augmented reality. And we've done the stack so many times. I'm going to do it again because I haven't done it in months in the show. You need to have a chip that can process what a camera is seeing in real, time. You need to have connectivity to figure out what's happening like in the world. You need to have a battery that can power it all. You need to have a display that is bright enough outside to show you stuff overlaid on things that are themselves bright and has a large field of

view. These are each of them massive technological hurdles. Oh, and then you need a battery that can make that all go. Right. The Vision Pro is like, the battery is, like, the battery is, totally external. Right? One solution to that problem. I think the meta display doesn't solve most of those problems.

They've literally just added a display. Yeah. They don't have the compute to do anything other than you can turn right. They certainly don't have like real time camera recognition.

Then they're meta and I don't want them to have it. Yep. I don't want them to have it. I don't want a bunch of people wearing Zuckerberg glasses walking around like live streaming

back to meta's data centers. It just seems to like as a society, we've all made the decision that that would be a bad idea. We're careening towards that future because no one can stop it. But like, I think anytime we discuss this product, we have to look at it. Then I am very concerned that in a world of like no content moderation, everyone will live in like differently augmented realities. Like right now, if you stare at the capital building, you're like, what happened here on January 6th? Like, different people have wildly different opinions

when happened in that building on January 6th. Meta will have to make some decisions about what to say, and then Donald Trump might throw Mark Zuckerberg in jail if he doesn't say the right things. Like, that's just a weird place to be. It's a set of technology, a set of experiences, and they're, like, so excited about it. I'm like, have you guys...

You know, it doesn't... A, the tech to do it doesn't exist, and then that outcome is careening towards us as well. I mean, I would say, like, to the implicit question there of, have you thought this all the way through, the answer is like, of course not. What about the history of the tech industry suggests that anyone has thought this all the way through? We can't convince Open AI to think, gosh, how will copyright owners feel about it? Like, this is just not what we do here. We ship the thing.

See what horrible stuff happens. Apologize. That's what we do. It's the tech industry. But I also think a bunch of that is really far away. Right? And so the question is what's the next thing? Like, the thing you just described doesn't work yet. And it won't work anytime soon. The like, can I stand in front of the Capitol building and be like, what happened here? And it will like usefully tell me how, like, none of that is real. It might be eventually. But even even the thing that is good and interesting and cool between here and there, I think we have not yet figured out. But I do think the idea of what is this going to look like on my face? We all pretty much know it's going to look more like meta than like. Division Pro. Yeah, I'll give you that. I mean, you know, the killer app is a killer app. If they can just do names and faces, everyone will buy one. Yes. I can buy that one. Yeah. Mark Zuckerberg,

all is forgiven if you can let me do names and faces. And I know they can because the prototypes have been developed and the kids are building proofs of concept of this using the existing meta display app. And then they got a lot of publicity in those kids, I believe, started a company to do it for real. Perfect. Real story. There you go. So like that's coming. And I, you know, set of questions is going to come along with that very, very fast. Yeah, yeah. Right. We actually have some more AI stuff to talk about, but we should take a break first.

And then we're going to come back and like you, you have just been alone with AI tools for three months and we need to talk about it. I have a new girlfriend, David. This is the last time I'll ever see any line. But we'll get into that. We'll be right back. Support for the show comes from Framer. Framer is an enterprise-grade no-code website builder used by teams at companies like Perplexity and Muro to move faster.

With real-time collaboration and a robust CMS, with everything you need for great SEO, not to mention advanced analytics that include integrated A-B testing, your designers and marketers are empowered to build and maximize your dot-com from day one. So whether you want to launch a new site, test a few landing pages, or migrate your full.com,

Framer has programs for startups, scale-ups, and large enterprises to make going from idea to live site as easy and fast as possible. Learn how you can get more out of your dot com

from a Framer specialist or get started building for free today at Framer.com slash verge for 30% off a Framer Pro annual plan. That's Framer.com slash verge for 30% off. Framer.com slash verge.

Rules and restrictions may apply. Support for the show comes from Grammarly. You don't need reminding that the world moves fast. But work today requires clear communication, and when every message counts, sounding rushed or generic can be getting lost in the shuffle. Grammally gives you one place to think,

write, and finish your work where you already write, while giving you access to agents that help you sound natural and engaging. No matter what kind of, of writing you're doing, Grammarly helps you get ideas done faster and move from draft to done with less friction. You can use Grammali's AI chat to brainstorm ideas, outline a solid draft, then refine it with context-aware suggestions that fit what you're working on. See why 90% of

professionals say Gramerly has saved them time writing and editing their work. In a world of generic AI, you don't have to sound like everyone else. With Gramerly, you never will. Download Grammarly for free at Grammarly.com. That's Grammarly.com. Support for the show comes from LinkedIn. If you're a small business owner, you know that every hire counts, but time and resources are limited. Finding, connecting with, and screening the right candidates takes up valuable time you

could be giving to your customers. That's where LinkedIn Hiring Pro comes in. It's built to be your hiring partner, helping you find the right candidates. faster. That way you can hire with confidence without turning it into another full-time job. Hiring Pro streamlines the entire process from drafting your job to shortlisting candidates and conducting AI-powered interviews for initial screenings. Its updated conversational interface lets you describe what you need in plain language. Nearly 60% of hires find a candidate to interview within a week. With Hiring Pro, you spend less time searching and more time connecting

with the right talent. And instead of getting buried in resumes, you get a focus shortlist that actually moves your hiring forward. Join the 2.7 million small businesses using LinkedIn to hire. Get started by posting your job for free at LinkedIn.com slash track.

Terms and conditions apply. All right, we're back. Neelai, when I came back, I spent two weeks shouting my insane sleep-deprived takes at Jake and Hayden, and I offered you the chance to do the same here.

And you said, basically, you only had one and it was about AI. Yeah. Do you have something you want to get off your chest here? So I have the same basic AI take as I think a lot of people.

I'm going to say that my path to getting there is the virgious path. Okay. So I think we're in a bubble. I think we're in a big AI bubble and it will eventually pop. And, you know,

I agree that A. Open AI doing sex bots is like pretty bubbly. They're just bubbly. There's a, you know, on the circular deals. Liz sent me a chart of all the circular deals of the companies investing in each other today.

And let me quote specifically what she said. She wrote, she drew it and sent me a picture for a story she's working on. And she wrote, hello, I made you something horrible. It's basically just a, it's like a series of circles with labels. It's like a lot. So that's like, well,

and I think that's like the standard argument, right? Yeah. The money is going in circles. That's not good. Well, yeah, the news this week was opening eyes like, we're going to spend a trillion dollars on compute. And like, and the Nvidia stock goes up, and AMD stock goes up. And everybody's stock goes up every time Sam Alton says, like a bigger number about computing costs.

And then somebody raises their hand and goes, where is that money going to come from? And Sam Altman just sort of like throws a smoke bomb and disappears and just like keeps getting away with this. And this, it continues to work. And everybody believes him about this money for now. But the we're in a bubble. talk has gone from sort of a thing people were saying quietly to just like mainstream accepted

really truly. It's been really interesting. And again, I think that most that is like financially oriented. Yeah. Right. This is a bubble because this finance stuff doesn't work. I heard Kara Swisher say the circular deal making reminded her of like telecom in the 90s. Hmm.

Great. There's that whole argument and you can get that anywhere. You can go buy that on the street corner. It's fine. This is a Vergecast. Here's my argument. LLMs as a technology cannot do all the things you need them to do

in order to make the products that would pay all this investment off. And a lot of people can argue with me about this. There's a lot of ways to supplement LLMs as a technology. There's a lot of ways to enhance them. There's a lot of ways to connect them to traditional deterministic computers. But the way that I know it can't be done

is I've watched the three biggest tech companies in the world, consumer tech companies in the world, Apple, Google, Amazon. all try to ship a smart home assistant based on an LLM, and they can't do it. They simply cannot turn the lights on and off, reliably, using the foundation of what modern AI is. I'm not saying there's not other AI techniques, right?

LMs aren't all of it. There's more. You can layer stuff around them. But everyone knows what it should be. Like, I know what would pay off all of this investment. Yes. It's the super capable assistant that can just do stuff when I ask it to do stuff. And I know this because we keep getting shown the demos of it, right?

Computer use from OpenAI, whatever the Chrome thing we saw at Iowa is where Chrome went and browsed for you. We've talked about whether the web would exist at all because the robots could just use the computers on your behalf. You'll never book a flight or order a sandwich again, right? Like, we've been shown, Microsoft announced that the next major version of Windows would just be a computer you could talk to,

all built around AI. I've heard this from Microsoft before. Yep. Many, many, many times. But this is the vision that will pay off the better. You will have an ambient computer around you. You will speak into existence your desires, and then it will happen.

Yes. Three of the biggest companies in the world cannot reliably manage to turn the lights on it off. Straightforwardly, like Gen 2E has reviewed Alexa Plus, and that, first of all, they're adding ads,

just like straight up ads to all those displays so that people get them out of here. They call it orchestration between when you talk to the AI and it, you know, figures out what system to use. None of that is reliable.

Nope. Sometimes it hallucinates and just lies to you. Sometimes it works, and sometimes it doesn't. The real promise of like, hey, when I get home, Like, every time I get home, like, detect that I'm home, turn on the lights, start the coffee maker, do this stuff where you're basically programming with prompts and you get automations.

Real shaky. Google is starting to roll it out, but it's just adding Gemini is like a feature to the existing Google assistant on these hubs. Right. They haven't even, like, directly connected them. And you have to pay 20 bucks a month. Yeah. Yeah. But I think that the Google answer is particularly telling because that is clearly the right answer, right? Like the sort of less impressive version of these systems, which is what they've been building since like 2012 with Google Assistant and Siri and Alexa and all these things, they're not interesting in the same way that a lot of the agentic AI stuff is interesting right now to people. But it just work. Like when you ask your assistant to turn off the light, it will much more reliably turn off the light if it's running the old tech than if

it's running the new tech. And that is kind of in a way, like it's a feature not a bug. And it's actually because these LLMs are not trained in such a way to just go from A to B like that. They are not, they are not command executors. That's not what they do. There's a bunch of steps in between, like you said, all of which can fall apart all the way along the chain. And the chain is big and interesting and open and allows you to do lots of things. But what people want to do is turn off the light. No, it's fascinating to me. And, you know, I've just added more and more smart home stuff to this house. You know, my wife and daughter are like, we're never leaving this house. That's just where they are. A seven-year-old girl, by the way, saying, we're never leaving this house is like,

her foot is down. It's, we're not going on. It's like, I can do it. Like, the house is a robot now. Like, we're fine. We're all in. And so I'm like, you know, we have a home assistant set up over here for me to tinker with. We've got a home bridge over there to bridge everything. together. And I'm like, oh, the way that this should work is I, you know, it's my network. I should put one of these assistants on my network. It should just see that there are some hugh lights over there and there's a TV over there and this thing. And it should just go use it for me, right? I should be able to give it credentials and just use it. I should actually have a smart home platform. That's how this should work. Like my house has a bunch of network

devices on it and this thing should, you know, assuming that I trust all these companies. have permissions to traverse my local network and flip switches for me. And nope. Like, just nope. Well, and less do you think that is like an outrageous thing to ask, Ero, the company that makes your Wi-Fi router, owned by Amazon, the company that makes Alexa.

Like, it's sitting there. It's all right there. Yeah. Right. And they just can't do it. I mean, Apple can't do it at all, right? They delayed Siri. Right.

And they control all the applications on your phone. And they couldn't do it. And I don't know if you guys talked about this. past few weeks, there are some big hints that Siri app intents, which was the thing they were going to make developers do so, you know, AI Siri could control the applications, that they were going to use MCP model context protocol, which is the sort of nascent protocol that Anthropic developed to do all this stuff, that Apple would like basically cave and give in and use that, which would be a huge cave. Yeah. Right. Because Apple not doing its own weird standard is usually a sign that something

has gone massively awry. Yeah. And, and it's a huge cave. And, you know, and it's a huge cave. And, you know, if you look at the Google subreddit, the Google Home Subreddit, everyone is furious at Google all the time. They're mad. They don't trust this company. They don't want this stuff. And it's like Google, you have this huge installed base of unhappy people. And for some reason, you're charging, what is it, 20 bucks a month to use Gemini on a nest. This is insane. You could just give them a good product that they like, and they would be happy. And you have lots of cuts of this product. The Gemini is all over the place, right? You can't open Chrome without it insisting that it's summarized whatever you're looking. It's actually. This is their installed base. I mean, you're, you feel trapped by the phone. You're Google and you're like all of our customers or iPhone users, like all of our most wealthy customers are iPhone users. You're like, we can't make another fire phone.

All of our customers are on someone else's platform. They have both given away millions of little voice-activated pucks. They have the AI chatbot interface sitting in millions upon millions of homes, and they cannot ship the product that would pay off that investment. And so to me, this is the canary in the coal mine. Like, it is, I've used a lot of AI tools. AI has been more useful than I anticipated it being in a lot of ways. Lutron, Caseta, wiring diagrams are very bad for those light switches.

And if you just ask Chat Chappi to help you, it will just help you. It will get it correct. And then it will produce a diagram that will burn down your home. Big disconnect there. In text, get it right. Photo, burn down your home. Like, I don't know what it's doing. But, you know, we'll talk you through stuff.

That's all very useful. But I just think the fact that no one can ship this product, the rabbit couldn't do it, the Humane Pin couldn't do it. I saw Johnny I at OpenAI Dev Day. And I think Casey Newton, like, posted, like, he's saying exactly what I would say about my forthcoming AI device if I had no idea what that device was. Yeah, I think it was Casey who posted something like he talked for five minutes about what an idea is and ended with, it's just an idea, you know.

I mean, it's just like, this is, I don't know if you feel the same way, but I'm just looking at what will pay this all off, what will make this not a bubble? And it is ambient computing, right? It is the dream. Yes, completely. Not even AGI. I'm just literally like, I'm in my house.

I'm like, can you go? We're out of oranges, like, order some more. And like, something happens. We're nowhere close. No. And I think, I've been trying to think a lot about the disconnect between the way that people who love AI talk about AI and the, way that people experience AI. And I think the cynical version of the AI disconnect is that everyone who

likes AI has a financial stake in AI being successful. And I think that's not wrong. Like, every time you hear anyone tell you why AI is terrific, you should just hear, hopefully this makes my stock price go up in parentheses at the end of every sentence. Like, that's, it explains everything. But I also think the thing that is true about LLMs is they make what people call fuzzy search work. Yeah. Right? Like this thing where I want access, I want to communicate something to you, but I don't have the exact right set of words in the right order to speak computer about it.

I just want to say something and have you understand it. It does understand it. Like that is the magical thing that happened with LLMs is I can just say a bunch of quasi-sensical words to my computer, and it will actually understand what I'm trying to say far more reliably than anything we've had before. That is hugely powerful.

Wait, can I refine that idea ever so slightly? Sure. It doesn't actually understand it. I think this is the fundamental problem. I think it will accept it as input and return reliably and consistently something useful. But the part where you're like,

turn on the light, and it has no action actual understanding of light or turning on or like success, right? Like, that's the gap. Maybe. There's a feedback loop in there that is just, that's why I think the smart home is the canary in the cold mind. Well, but I think there's two pieces into that, right?

One is, does it actually understand what you're saying, or is it just getting better at appearing to understand, right? And this is like, the thing I've been saying forever is that AI is getting convincing much more quickly than it's getting good. And I buy it. That's still true. But I do think there's something to, like this to-do list app I use called To-Doist. They put out this feature where you basically just like ramble your thoughts into it and it pulls

a to-do list out of it. And it's like, it's very good. Like you just sort of sit and talk at your phone for two minutes and it's like, here are the six tasks that you just said to me that you have to do today. That one step thing, it can do fairly well, right? It's text in, text out or texting image out even is like that's a thing that's slightly getting better. we so massively underrate the number of steps between me saying turn the light on and the light turning on

and just the sheer number of things that can go wrong between there. And even if somebody was saying this to me the other day, and I've been thinking about ever since, even if each step along the way works 95% of the time, you get a few steps in and you're at like a 50% failure. Yep. Yeah, you're just stacking up here. Like these things have to be at like five nines each, 99.999% success each every time before it's even going to feel like it works most of the time. And like you stretch that out to legitimately like it's probably like tens if not more steps in between I want to turn on the light and the light goes on in my house with the way that LMs are set up now. And I just I have precisely no faith that it's ever going to work. And if it works six out of ten times, it's an abject failure of a product.

So I mean, I think that the error handling for most people is, like, understanding that the light isn't on. Sure. You know what I mean? Like, you're describing a chain of events that has to go well. And if you actually had an intelligence operating that chain of events, they would all go well. Because when they went badly, the intelligence would say, well, that is not going the way I wanted it to. Right. I'm going to try something else.

I'm going to get to the outcome. Instead, you have like, that didn't go the way I wanted to, but that's a great question. Have you thought about this? And it's like, what is going on here? And I'm not saying that's how it works. And I know that there are in particularly like enterprise software deployments. There are agentic systems that work well. And boy, do AI people love posting me about that when I complain about this.

But my point is, if you put the stuff on Rails inside a really well-known data sets with very, very defined tasks, you can get very impressive results. Like, cloud code is a very impressive product, like straightforwardly. Sure. And it can do a bunch of agentic stuff. They announced today, Skopax, where you just like, basically, the way they describe it to me always sounds like the Matrix. They're like, you just give it some capabilities.

Now it's good at Excel. Like, that's cool, right? That's also just software. Like, to be so clear, that's just software. That's not magical, artificially intelligent, anything. that is just you built software. This is the kind of stuff people have been making software to do

for decades. Right. And the thing you are specifically describing is, oh, now we have this very powerful, very convincing natural language input system for the software. Great.

That is one part of the dream. But it is just like the AR glasses tech stack. Right? In order to build this thing, you need a very powerful natural language input system that can accept all kinds of input

with no formatting and no syntax. and all this stuff, and it will be convincingly be like, I know what to do now. Then you need a bunch of stuff that can do it. Right. And, you know, the idea that this thing,

it will linearly scale to be a super intelligence, just feels wrong. Like, based on the products that I, you know, used over this break, based on the fact that, again, the biggest companies in the world

have every possible incentive to make this happen. Yes. Apple does not want to see the iPhone disrupted with Johnny Ives' voice-activated idea. they don't. Humane for all this disaster flaws was like,

oh, we can, if we get this right, we can beat the phone. Right. And they couldn't make it. Right. Rabbit, got all that hype at CS because it was an obviously new input method. Google should be like every Google Nest product

should be the thing that displaces the iPhone. If you're at home, you should not be using your phone except to watch TikTok or whatever. Like all of your questions and ideas that, right, we have a big antitrust trial. We have this weird deal with Apple that is always under risk. Who knows what's going to happen here? We should just move you to the nest.

And they can't do it. Amazon, Panos-Pen-A. I know he wants to do it. We've talked to him about it. He wants to do it. It's harder than it looks. And I, to me, that is the canary in the coal mine. Like, everyone else can talk about the financing and trading stock for computer,

whatever that is. So it's all very abstract. I'm like, does it work? Can this go from here to there? Is there, can you show me how it will get from here to there? Like, you could show me Bluetooth, which it took a long time. Yeah. But every year they'd be like, this version of Bluetooth gets us closer to AirPods, and then there were AirPods.

And everyone knew from the very beginning that AirPods was a product they wanted to make. Yes. There are two arguments in this space that I am so deeply tired of that I still hear all the time. The first one is, well, this is the worst it's ever going to be. And it's what everybody says about everything, right? Like the new image generators come out and people are like using them and finding the edges of them and things breaking it weird. People are like, well, this is the worst it's ever going to be. Like Elon Musk when Sora came out and they're doing this stuff in Grock was talking about, you know,

there's going to be a feature film made by AI next year and the year after that it's going to be good. And it's like, A, that's a very funny admission. But B is like everyone assumes, like you said, that we are on this. unstoppable linear path to all of this is going to keep getting better at exactly the same rate forever and ever and ever. And like, nothing works like that. There's just, there's just no version of the history that suggests that that is how it will go. And already, actually, there are lots and lots of signs that we are hitting plateaus really quickly, that, that the, like, marginal gain in making your model vastly bigger is slowing down. Uh,

That's the one argument. And then the other argument is that it is to show me a chart of how fast things are getting and how much better they are and just keep drawing the chart. Yeah. It's like, again, it's the linear thing where it's like, well, it was here and now it's here. So by the obvious transitive property of math, it's going to be up here next year. And again, I'm like, all of you are too smart to actually think this is how this works. that like we're just going to keep marching on the path of all of this getting better. And so the only thing we have to do is keep doing what we're doing and eventually all of these

problems will be solved. But then when I ask people questions like yours where it's like, okay, how are you going to, how are you going to like correct errors? Right? The opening I put out this really interesting paper a few weeks ago now, I think, about why models hallucinate. And one of these people were talking about around that paper was basically one really interesting piece of research is how do you teach a model to understand when it's gone wrong, go back and start over. Right? Because that's what it needs to do.

You have to teach this thing like, oh, I've gone awry. Let me pull this back and try a new thing. Because one of the reasons they hallucinate is they make one small mistake and then they compound that small mistake over and over and over until they made a very large mistake, right? Like that's how it starts. And so what you have to do is get this thing to go, okay, I didn't turn on the light yet. Where did I go wrong? Yeah.

And the technology we have just does not support that now. They just can't do it. And the idea that we are just going to inevitably march in that direction just drives me a little tiny bit crazier every single day because it's just not happening. Like it's not like the products you're describing are getting slowly better. They just suck. Well, in the case of Apple, they don't exist.

Right? They just had to give up and appear to have started over in some major way. This is why I brought up the thing about App Intents and MCP, if you're starting over and you're hinting that you're going to use someone else's standard, that you have started over. Yes. Well, and Apple continues to sign up, you know, different partners for different parts of its. Like, Apple may just be fully walking away from trying to do it itself in any kind of meaningful way here.

Again, you know, maybe Google and Amazon will figure out. Let me just read you this quote based on what you just said. Let me just read the quote from the AIPC announcement that Microsoft made. This is Yusef Medi, who's the EVP in, chief marketing officer at Microsoft, basically in charge Windows. Here's a quote. You should be able to talk to your PC,

have it understand you, and then be able to have magic happen from that. That's huge, right? Like, that is legitimately Star Trek computer. That's also nothing. Like, he has no idea what this is for. Well, no, with your permission, you want to share with their AI in Windows

what you're doing, what they're seeing, the PC should be able to act on your behalf. the word should by the way this is another thing in the AI world the words words like should

and ought to and like these things do a lot of work because again it's what you're saying at the very beginning of this we know what this

is supposed to look like this is one of the things that's so enticing about AI it is so obvious how this stuff is supposed to work and what it would be like

for it to feel amazing is so clear and it's so enticing because this is like there are pieces of this that seem to enable pieces of that that are very difficult. But, like, yes, you should be able

to talk to your PC, have it understand you, and then have magic happen from that. Yusuf METI. Like, you're correct about that. I agree. 100%. Do you know how they're advertising Windows 11 right now? Giant tagline, the tagline for these PCs in Windows 11 is meet the computer you can talk to. So they're not, you should

be able to do this. We are advertising the PC as a thing you can talk to. And you just see, like, we're just, we're We're not going to meet people where they are. We're going to tell them they can do something and it won't do it. Windows. Like, what do you want?

Yeah. And there's just this, you know, you talk to regular people who are using AI. I have a number of friends now who just refer to chat. Chabit is Chad. They started saying chat and they slid their way to Chad. Wait, that's really good. It's pretty good. I actually really like that.

And I just use chat for everything. And I'm like, you are going to crash your car, man. I'm like, what do you want me to say? Also, Chad, GBT being Chad, makes all of Chad's mistakes so much easier to deal with. It's like, oh, it's just Chad being Chad, you know what I mean? But that's my argument is like, just like I think Matt is confused about the glasses really being a camera.

I think everyone is looking at people hanging out with Chad and being like, this should do everything. Yes. And maybe it should. Maybe you should just walk up to your Windows computer and be like, make these edits and these photos and I'll just figure, like, that would be great. millions of miles away from this.

Well, and again, the idea that it is fun to make SpongeBob SquarePants do meth with an image generator and I want to use this for work all day, every day, like those are so not the same thing I can't even describe it. And everybody, there is this belief that it's like, it's going to be like the smartphone, right? Whereas if you want to do one thing on it, eventually you're going to want to do everything this way.

It's like, I don't think so. I just, we've been trying to make people talk to their computers for a really long time. And one reason they haven't is because it hasn't been very good. But the other reason is just that I don't think people really want to sit there and like yammer away at their computers all day. Yeah. Like, for me, I don't know what your experience is like, but I talk to my assistant all the time in very specific use cases. I do it in the car to like set reminders. I'll do it to like take notes.

I like to do voice messages sometimes when I'm just like out walking. It's just like an easier thing than holding my phone. There has not been one time where I'm sitting in front of a keyboard and have chosen instead to talk to my device. Like maybe I'm just old, but like that, that I have not yet encountered a reason for that behavior to exist. I mean, we have a Nestub. I mean, I have one here in this office and it is just a photo frame. I never talk to it when I'm using my actual computer. I don't need to.

It's actually quicker to play music by clicking on the Spotify app. my computer. Like, it's, I don't, I don't know. The new car has Google Assistant in it. Max loves it because she can just ask for Taylor Swift. She was very sad about the Jeep going away, but the Jeep has since been recalled yet again. So the car needed to go. It's just a kid out of my face.

This time I got recalled so hard that it loses propulsion. Oh, yeah, I heard about this. This is just a U-Connect update, like an infotainment update on that Jeep loses propulsion. Not great. It needed to get on my face. Tax credit's going away. Bought an EV, got a great deal. Doing great.

Max is sad about the car because shade strange. And I was like, you can just ask Google to play Taylor Swift. Car is the best car we've ever had. We're never getting rid of this car. Is it like full Android Auto gas stuff? Oh, yeah. It's GM. No car play, the whole thing.

We have some Decoder stuff coming up with GM. Joanna Stern, by the way, did a great job yelling at people about car play on Decoder the last few weeks. She did. People should go listen to those. They were good episodes. She got a lot of mileage out of my bit. You know what I mean? She did.

No, she did great. All the guests didn't go to a great job, but it was particularly fun of Joanna because she basically live texted me all of her ideas. It was good. And I got to ask Ford CEO about a very nerdy ECU question, and he threw his own daughter's boyfriend under the bus choice. Very good.

Go listen to the end of that episode. Very good. Anyhow, I asked Google Assistant to open and close the garage because the garage door button in the car is on the screen. It's not a hard button. Google Assistant When you do that, I'm reasonably sure

it goes through the smart home system, not the garage door in the car. Right. Because it sometimes asks me for a pincode, which if I'm already in the car looking at the button, I should not need to enter or say out loud a pin code to open the garage.

I'm in the car, I'm looking at the button. I can just reach out and push the button. And it's like that kind of disconnect all over the place, or even in the place where it should work the best. You can see, oh, these are multiple systems that are just trying various things to get the job done instead of the most obvious one. And the more companies roll these out bad, I think the harder they're going to make it to really

make this stuff mainstream. Like, I don't think there are going to be legions of people out there who give these things infinite tries. That I think, like, Chad GPT, a lot of these things, Chad GPDC, Claude, the, Character AI, all these things get away with a lot because they're fun to talk to. But the things that are sort of deliberately trying to be useful, people are just going to bail. And it's what we saw with the first generation of the smart home stuff, that people just stopped caring because it wasn't worth the hassle.

I can just walk over and turn off the light. And like, would I rather have my assistant do it for me? Sure. But I've asked it three times and it didn't do it. And I'm going to stop banging my head against this particular wall. I'm just going to walk over and turn off the lights. and I think every one of these companies runs the risk of giving people that same feeling.

Like the hill Apple has to climb, even if Siri's amazing, the hill Apple has to climb to get people to use it again at scale in a meaningful way is so, so huge. And I don't think you get three tries at this. Like if new Siri sucks, they better stop calling it Siri. Yeah, and there's all these rumors this week

that Apple's going to do a bunch of in-home displays, including one that swivels around and one you mount on your wall and they'll have cameras and it's like, yeah, if you put a thing in someone's house and you can talk to it, the expectation now is it's chat chitb t level.

Yeah. And again, you can build that product. You sure can. You can put an iPad in your living room and put the chat chabit app on it and you can talk to it all day and not. And then no one can make the next turn. I'm just saying Canary and the coal mine.

That's my hottest take. Like, you can do the financial bubble stuff. I'm saying lights, light switches, canary in the coal mine. That's the bubble. It's a good one. All right, one more thing on the chat GPT front before we go. We should just talk about Sam Altman sexting with chat GPT. So the basic news this week was that opening eyes planning to roll back some of its restrictions

on the way that chat GPT has worked to allow more people, especially adults, to do more things with chat GPT. There's this very funny thing that Sam keeps saying that like the main criticism open AI gets is that there are too many guardrails on chat GPT. And I'm just like, you should maybe like vet the people who are giving you this particular kind of feedback a little more. But this is this is a thing that they've been saying for a while. And they said, I think a few weeks ago that they were going to start to allow people to make, quote, mature chat GPT apps.

But basically, Sam said specifically they're going to allow people to make erotica with chat GPT. people had a lot of thoughts about this. And then Sam went on X to clarify this. This was, I believe, Wednesday. And he gave this long response about how, for young people, particularly teenagers, they're prioritizing safety over privacy and freedom.

That was the phrase he used. And that they're also trying to be very careful about mental health. But then he said, and it's just like this, see if this resonates in your brain and reminds you of the same thing he did with me. But we are not the elected moral police of the world. In the same way that society differentiates other appropriate boundaries, R-rated movies, for example, we want to do a similar thing here.

And that's how he ends. Just a full, full handwash. You do whatever you want. It's not my problem. Yeah, technology is a tool, and we're not responsible for how people use it. Pass a law. Said Mark Zuckerberg in 2011. There it is.

It's just like, we're just doing that again. Yeah. I mean, look, they need, they can't. they can't let anyone else do AI girlfriends. What's the stickiest thing they got? Completely. AI girlfriends. Yeah.

For better or worse, like weird, that weirdness is sticky for all of these apps. I don't know if anybody ever wants to date co-pilot, but you will recall AI burst onto the scene when Microsoft bang, tried to bang our friend Kevin Roos on the front page of the New York Times. I hope people have moved on. They have read reports that some people have not. They super haven't moved on. I hate to tell you, you went away and everybody just kept banging chat GPT. Like, I don't know, man. Does Sam Haldman want a fuck a robot?

He sure does. Like, it's very clear to me that they're like, I want to bang a row. I want to bang my laptop as an animating force in society. I, we are not the elected moral police of the world is like, yeah, but you should have taste in your products. Like, on the other side of this, which in a way that I thought was really out of bounds at the time, Steve Jobs used to say all the time, there's stuff I'm not allowing in the app store. Yeah. We are not going to allow erotic in the app store.

Remember, Comicsology, it used to be in a big fight with Steve Jobs. That's right. Over, like, what the content of the comic books and the comicsology app was. And then they would often point to the fact that HBO Go was on the app store, and it had just, like, straight up HBO softcore porn in it. And he, like, yeah, but they're HBO. Like, they get to do wherever they want. Like, Steve Jobs set those boundaries on the app store.

And he would often say people have to trust us. The web is available to you. I think Sam, they need to make money. Like, that's what this looks like to me. They have to make money. They have to be sticky. They have to have people using it all the time. They got to, those data centers need to be running red hot.

Otherwise, none of this investment is worth it. And like, you can't have people go to replica instead of chat chubby T. Yeah. I mean, and it's like, it's a truism of the internet, right, that all the money is secretly in porn. Like, we talk about everything else, but all the traffic and innovation and money is actually in various kinds of porn. If you're Sam Altman and you need to give chip companies a trillion dollars, like one really good way to do it is to chase this thing that is clearly going to be popular and clearly going to be a big business. And rather than embrace it, just wash your hands of it. Right? Like, it's a, I would almost be more into this if he was just like, we're launching porn GPT, let's go.

Like, I don't think that's a great idea, but I would actually appreciate that. I would actually appreciate the candor of porn GPT more than to just say, well, we're not the moral police. We've just created this tool that we don't understand and can't figure out how to put rails on. So we're just taking them all off. Do what you want. Yeah, this is Marks. I'm going to be the arbiters of truth. Yeah, and this is how you stop doing content moderation on your platform. Have you opened Facebook lately? Like just scrolled through the average blue app feed.

It's awful. It is awash in AI nonsense. Yes. Basketball players giving quotes about things that they've obviously never considered before in their life. Any number of NFL players, AI quotes about Bad Bunny, right, all day long. There was a great one going around the other day that was, because there's been, for anyone who doesn't follow the NFL. Patrick Mahomes, very good quarterback. there's a belief among people

who don't like Patrick Mahomes that he gets all the calls from the referees and there's this great clip going around where somebody was like That's not a belief. That's statistical fact. There's this great video going around

that Patrick Mahomes walks up to the ref he like, you know, daps him up and then he gives him a hug and then they just do like a light, gentle kiss through his helmet. And then he walks away. And it's like, and it's just one of those things

that it like, it goes around just enough that it's like, oh my God, there are a bunch of people who watch this and are like, see? And that's the stuff that's all over my Facebook feed now. Yeah.

And it's because I was looking at it. But like there's like we're going to make porn GPT. There's, we're going to, there's Facebook saying we're not the arbiters of truth. And then there's this specific collision. We're not going to police what you make with the tools. We're not going to police what you distribute with the tools.

We're not going to try to tell anybody what's true or fake. And then you end up with these feeds full of slop that is ever more convincing. And truly, like the number, it's beyond. Like, is this AI or this is very funny? It's just my feat is people just reacting to obviously fake things as though they are true. Yes, 100%. All day long.

The specific one is, quote, fake quotes from NFL players about bad money and how they would never allow an anonymous. And it's like, and then the whole, and that's not even hard to make. That's just like with Photoshop. But like people just believe that's true. And now it's going to be video. Now it's going to be deep fake erotica of celebrities. I mean, the hop from SORA to doing that is zero. Oh, yeah.

And it's very clear that OpenAI is perfectly happy to let that happen because, you know, let adults be adults. Or, I mean, no, it's not even let adults be adults. If the people are dead, no one can stop them. There's that, too. Right? So, like, the number of weird Michael Jackson videos I've seen from SORA on Facebook, the people believe are real, is insane. Interesting.

That one I have not seen. Oh, it's horrifying. That would freak me out. Yeah, I don't like that at all. Okay, please do. All right. In the meantime, we should take a break. Sam.

We're going to go look at some Michael Jackson videos. And then we're going to come back and do the lightning round. We'll be right back. Support for the show comes from MongoDB. If you're tired of database limitations and architectures that break when you scale, it's time to think outside of rows and columns. Because let's be honest, you didn't get into tech to babysit a broken database. You got into it to actually.

build something. MongoDB lets you do that. It's flexible, developer first, acid compliant, enterprise ready, and built for the AI era. Say goodbye to bottlenecks and legacy code. Start innovating with MongoDB. There's a reason it's trusted by so many of the Fortune 500. And that's because it's a platform built by developers for developers. MongoDB. It's a great freaking database. Start building at MongoDB.com slash build. Support for the show comes from LinkedIn. If you're a small business owner, you know that every hire counts, but time and resources are limited. Finding, connecting with, and screening the right candidates takes up valuable time you could be giving to your customers. That's where LinkedIn Hiring Pro comes in. It's built to be your hiring

partner, helping you find the right candidates faster. That way you can hire with confidence without turning it into another full-time job. Hiring Pro streamlines the entire process from drafting your job to shortlisting candidates and conducting AI-powered interviews for initial screenings. Its updated conversational interface lets you describe what you need in plain language. Nearly 60% of hirers find a candidate to interview within a week. With Hiring Pro, you spend less than searching and more time connecting with the right talent. And instead of getting buried in resumes, you get a focus shortlist that actually moves your hiring forward.

Join the 2.7 million small businesses using LinkedIn to hire. Get started by posting your job for free at LinkedIn.com slash track. Terms and conditions apply. Buzzwords like progressive and affordability are thrown around all the time in politics. But what do they actually mean? For me, being a progressive means at least two things. One, being willing to unite lots and lots of people, all of the folks that are getting screwed over

against the powers that be that are making your life worse. And then second, being progressive is essentially a hopeful enterprise. That you think, I think that the world can be much better, that we don't have to settle for crumbs or settle for the status quo. And is there a difference between what it means? to the elected officials and what it means to the people. So money is essentially the root of everything. I don't care if you're gay.

I don't care if you have all that. That's like secondary, third. Like that doesn't, that's not a priority. That's this week on America Actually. Let's begin. All right. We're back. It's time for the lightning round.

Unsponsored. Niel, you should say it. It feels better when you say it. I like it better. Sliding round. Unsponsored for flavor. There we go. We're so back.

We're working on ways to do this. Before I left, I think I said what we wanted was on staff influencers who would do the things that you couldn't pay us to do. We've got some ideas. That's all I'm going to say. We've got some ideas. Somebody just pops up over Nielai's face. It's just that guy in the corner of the React videos. Yeah, exactly.

They're just like waving liquor. Yeah, they just do the link right here thing. And then the link appears. It's good. We got a lot of things to do. All right. So you've been gone a while. So you should know when we left, Jake made the lightning round.

the Thunder round. I have since killed Jake, and I have since killed the Thunder Round. It is once again the Lightning Round, but the difference is we still try to do a lot of things in the Lightning Round instead of what you and I normally do, which is just like ramble nonsensically about one story and do more show. Yeah. So we're going to do the Lightning Round. So the way this works is we each pick a few things and we try to just blow through them as fast as we can. We're going to do three each. You're newly back. You get to go first. All right. Mine is a TikTok that is a representative of a genre of TikToks and a genre of music, I would say. And I've been so anti-AI that I want to preface this by saying the first time I saw one of these,

I immediately sent it to our friend Charlie Harding from Switched on Pop. And all I said was, oh, no, I think AI is good now. All right. Here's the TikTok. The song's kind of a bop, dude. It's good. So what you're listening to is an AI-created soul music. cover of Bulls on Parade by Rage Against the Machine.

It's a lot. This is making, so I, a thing I see a lot of on TikTok is like song, but what if something. For me, it's like mostly pop punk covers of various songs, which should tell you an awful lot about my personality. But I would not have clocked this video as AI, straight up would not have. And it is like, kind of insane. I'm very worried about you, David, because I believe it says AI generated and the guy who made it is even saying AI generated.

He is, I would say, the tip of the sphere of this genre. Like, you can just go get his album on Spotify. It's all AI generated soul covers. Nothing against the pop punk community. I would say that I'm a tangential member, a more of an old school 76 kind of punk guy. But making a pop-pump cover of anything is like fairly easy. If you're in a pop-bunk band,

you can make a pop-punk cover of anything. Play it louder and yell sometimes. You're like, I turn up the distortion. Making a soul cover. of a rage song is not easy, I would say, if you don't have access to a horn section. So AI is just like enabled, in particular,

old school soul covers of everything. They are all over the place. They're all over Spotify. These creators are making albums. They are the most fascinating re-contextualizations of music that I've heard in some time. I have many, many conflicted thoughts about cultural appropriation,

about the fact that in particular it is soul music that is being repurposed in this way. And then I'm like, these are all jams. And I don't know how to wrap any of that up. I don't know who's getting paid, right? The songwriters, by all rights, should be getting paid for this. Because they still have the writing credits of the song.

You can cover any song you want, and there's all kinds of licensing rules and say anybody can cover a song, but you still have to get paid. But I don't know if any of our automated systems understand that like Zach Dela Roker gets 10 cents whenever you play this on TikTok. Like I don't know.

Like there's so many questions. But I looked at this. I read all the comments which are as conflicted about AI, particularly making AI covers of rage against machine songs. And it's like, oh, this is the thing that has the people the most. They're like, this is really good, but I hate it. It's the, it's really good part that does it too. Because it's like it would be less alarming if it sucked.

Right, I think about the like Will Smith eating spaghetti stuff from a couple of years ago. It was like, that was weird and sort of made you feel gross, but it wasn't any good. Now it's like, oh, how do I feel about the fact that I like this song? The other thing is I have no idea how this was made. You cannot go into Suno or UDO and be like, here's Rage Against Machine Song, make a sole version. Like, it doesn't work that way. And there's some arrangements in here that are better than what the tools are capable of. So Charlie and I are actually, we're digging into it because we're both kind of fascinated by all of this.

He's digging the money. We'll come back to T.L. With some of results of these investigations because I'm dying to know how these creators are making these tracks, how the platforms are dealing with content idea and copyright and all this stuff. And like, there's something here that is more fascinating when it comes to AI and what we generate with AI than like the slop economy. But it's also like, oh, no. But you just see this stuff hit. And in particular because it's apparently, because it's AI, it's easy to do. Once it started, everyone started doing it and it started a snowball.

It's everywhere now, yeah. Which is like the way a dance trend works. But you don't get that with soul music? Like, that's weird. Fascinating. Have you seen the YouTube trend of Taylor Swift covering everything? That's another fascinating one. No.

But anyway, we should keep moving on. Charlie, we'll have Charlie come on the show and actually, help us resolve this. We love Charlie. It's also been a minute since Charlie's been on the show. So we'll have to do that. It'll be good one. All right.

My first one is it's always nice when streaming companies do silly name things for us. Like HBO Max became HBO Max again right before I came back, which I thought was really kind because they're like, look, we're going to undo our incredibly stupid branding decision that you've been talking about for two years. Welcome back, David. Apple, for you, decided to go from being Apple TV Plus. to just being Apple TV. And I'm renaming my son Plus now.

We love this for him. This now means you can watch Apple TV through Apple TV on your Apple TV, and that is a perfectly grammatically correct sentence. I just want everyone to know that. But also, this is fine, right? Like, this is a correct, normal thing to do. Just call it Apple TV.

But my favorite part of this was Eddie Q, who runs all of the services at Apple went on the town, Matt Bellany's very good podcast about Hollywood, and I would say basically described this decision as we just felt like it one day, so we did it. Yeah. And the quote is, we stayed consistent because of it, because he's talking about iCloud Plus and News Plus and the other pluses that Apple had. So he said, we stayed consistent because of it, but we all called it Apple TV, and we said, given where we are today, it's a great time to do it, so let's just do it. This is A, the kind of thing you can only say when you have like the ninth place streaming service that no one cares

about. And B is so transparently what it actually has been called. Like I would say most people in my life just call it Apple. I saw it on it's on Apple. Sure. That's that's the that's the distance people will go to describe this streaming service. Sure. Great. This is now the most monstrously named thing at Apple. They have three things called the Apple TV. Wait. And they're different things. So, well, first of all, everyone should go listen to this episode of The Town. I sent it to David this morning, and I was like, this is two people who just do not give a fuck. Like, it is the single best interview of an Apple executive I've ever heard. Because Matt Bellany doesn't care at all.

Yeah, Matt Bellany is like, I know everybody in Hollywood, and here's what they say about you. And it's not good. And then Eddie Q is like, I don't get a shit. Like, it's like amazing. He's like, I'm on the board of Ferrari. Like, whatever he wants is fine. Matt Bellany came on this show and basically was like Eddie Q just likes having famous friends.

Yeah, that's great. I mean, but I've listened to a lot of Apple executive interviews. People always ask me like, don't you want to get Tim Cook on the God? I'm like, no. Like, I would just fall asleep. Like, they're all robots, except for Eddie Q in this one specific interview where he just clearly doesn't give a shit. Like, I don't know how else to describe it.

And then I don't think it's about Apple TV. being the ninth place streaming service. I really don't. I think it's about the fact that the Apple TV 4K hardware is nowhere. It is just not relevant to people. True.

And so if your Apple, you have long since operated in a frame where your hardware is the most important thing and the software in the app is the second most important thing and the iPhone is the primary thing and everything else. It makes sense.

And I think they realized more people have the Apple TV app on their Roku just for the, the Apple TV content, then Apple TV 4Ks exist in the world. Oh, 100%. If anything, the box should be the Apple TV Plus.

Like, that would actually make more naming sense. Right, because you start with the main thing, and then the plus is the additional thing. Right. And there's no free Apple service. Like Apple News, if you don't pay for it, there's still stuff in there. ICloud, if you don't pay for it,

you get five gigs of storage, and then they hold a gun to your head until you pay more money. But like Apple, Like, I think they've just conceded the hardware race. So, like, yeah, we have an app on Roku's and Firesticks. Yeah. That's Apple TV.

It's this application, and it contains this content, and we're just... The idea that it's plus makes no sense to anyone. Anyway, I'm thrilled because I no longer have to remember whether our style guide allows me to do the plus sign or I have to write out plus. So this is a huge win for all of us here at the verge. What's your next one? Speaking of TV hardware, Tivo has sold its last deal. DVR.

Hmm, TiVo? Did you know that they were still selling DVRs? I am pretty sure it wasn't actually selling any to consumers that they were still selling them to like small cable systems. Oh, sure. Because that's what it's been for a while, right? It's like your cable box is powered by TiVo. That's kind of there.

Yeah, yeah. Like there's a whole thing. TiVo and its partners, their statement is our remaining inventory is now depleted. it's because the cable providers are all streaming providers now. They're not shipping hardware. I want those to be my last words before I die. My remaining inventory is now depleted. And then I just go.

But the reality is like this wasn't even their business. Their business was being a horrible patent troll for the longest time. And they stopped innovating a long time. I know an old TiVo product manager and he, the last time we wrote a TiVo story, he messaged me. It was like, why are you even paying attention to this? Like, they're literally just patent trolls. Like, they have no new. products, they have no new ideas. Like, they are there to collect revenue from people who do because

they invented a bunch of stuff in the 90s. So, wonderful. Congratulations, TiVo. It does feel like we should do a version history on TiVo at some point in the future. That's a good idea. Because now they're dead. Yeah. That story is over now. To extract rent from the American economy. And there was like, you know, the kids don't remember, but like peak TiVo was was a pretty real and powerful cultural phenomenon. Oh, TiVo now lives on as a software provider for smart TVs and vehicle infotainment systems. Oh,

the Blackberry strategy. No. We don't make anything. Do you want our software? No. RIP. Tivo, still a great name, good brand. Yeah.

RIP. I look forward to somebody buying it and bringing it back. Like, Napster's back. We're just, we'll do all of this again in 10 years, and I look forward to it. It's kind of a good. My next one, we'll stay on the TV thing for a minute. Roku had a bunch of announcements this week, and I would say I'm pretty loudly on record as saying,

I think Roku is bad now. I think there was a minute where... I think Roku made the best products for a long time, and then got very popular and decided it was an advertising company and just sent all of its hardware and software people on vacation. And that happened like six years ago, and they never came back. It's the only thing I can think of. But among their announcements this week, two things that I like very much.

One is there's a new upgrade to the voice assistant that they have on some of their devices where you can ask questions about whatever you're watching. I don't have a ton of faith that it will work super well. But it's like the little things that Amazon has done, like with X-ray where you can pull up the thing on the screen, it'll show you like who's in there and what's going on. That stuff is great. And it is this sort of little bit of ancillary content that I think is actually really useful in some of these.

and Roku is smart to do that kind of thing. We're like, one of the examples they give is like, what kind of fish is Nemo? Which is like actually the sort of thing that would be cool to do on my TV while I'm watching Finding Nemo. So kudos. But the other one is the Bluetooth headphone connectivity

that Roku has on some of its devices. The most innovative thing Roku ever did was private listening, which let you plug in a pair of headphones to the remote and just listen that way to your TV. Now you can connect Bluetooth headphones directly to some of the Roku devices. You can also use,

your phone app and headphones, which as the dad of a newborn, I have done for just dozens of hours over the last few months. Sitting on the couch, Roku app open on my phone, AirPods in my ears, kid asleep in my arms.

This is just like, that has been life recently. And it's fabulous. And now you can connect Bluetooth headphones directly to some of these things, including some of their cheaper like the streaming sticks.

You can actually just now connect Bluetooth headphones directly. Private listening is all. Awesome. Good job Roku. Make better products. I will add one to you, which is the AI on TV is like, Direct TV announced that it's going to start AI generating ads as screensavers on their hardware. It's not good. I don't like that. It's not good.

No. It's all coming. It's coming for Roku as well. And DirecTV's thing is like, it's, they're trying to do a like, we'll show you what you'll look like. like in these clothes in the ads? I don't like, I don't like any of it. No, no, no, no. Send direct TV a picture of your body. Do it, David.

For you, anything. All right, what's your last one? You know I had to end here. We got some Tesla cybertruck sales numbers. I wouldn't, I wouldn't say the wipers a hit, David. I would say that they only sold 5,385 cyber trucks in this last quarter. down 63% compared to last year, and they've only sold 16,000 cyber trucks so far in 2025, which is not 250,000, which is what Elon Musk predicted they would sell annually.

No, it isn't. That's solid observation from Neil I Patel. That is not, in fact, 250,000. Obviously, they're only going to sell 20,000 this year. That's the estimate, which would be one order of magnitude-less cyber trucks. They are sitting on lots. If you go to Tesla geoships and you know where they store their cars, you can just see acres of cyber trucks sitting around. Elon has started selling cyber trucks to his other companies.

So XAI has bought a bunch of cyber trucks recently. Why does XAI need cyber trucks? Because money has to move in a circle if you're an AI company. And that's how we're going to do it. That's the way it goes. That's true. Open AI is making circular deals. Elon Musk is making circular deals.

It's just like a much smaller circle. It's just him. He's like a ghost from the left back into the right one. And in the end, Grock is going to say some racist. Verge Superstar, Zoe Schiffer, as you know, you can check out of the verge, but you can never leave. She's at Wired now. She just did a piece where she interviewed Cybertruck owners at a cyber truck meetup in, I believe, Arizona. You should go read it because it's just a bunch of quotes.

One of the interviews has an editor's note, like, appended to it. It's like, this guy got arrested in January 6th and was later pardoned by Donald Trump. Wow. Choice. Totally choice. And then another one, perfect. She says to the guy, are you married? I was married, but I'm not anymore.

And then the only sentence that follows that is women do not like the cyber truck. Oh, man. All right. There is one of those in my neighborhood that I see driving around. And it was like, it was a very noticeable moment when the cyber truck got the like, I bought this before Elon went crazy bumper sticker and put it on. And I was just like, I know you own this bumper sticker. I also know when the cyber truck started shipping. Anyway, all right, I have one more, and then we're going to get it here.

Have you been paying attention to this, like, Pixel 10 Pro Fold drama? Oh, like Jerry Rig Everything, like folded it and it blew up or whatever? Yes. So Jerry Rig Everything, one of the great YouTube channel ideas of all time. They do a really good job of just, like, beating phones to death and reporting on them. I enjoy it very much. Can I just play you the chunk of this video in which he destroys the Pixel 10 ProFold? And I just want to know if you think this is a thing foldable phone owner should be alarmed by or not.

Okay. I'm just going to, I just want to play you this chunk of this video. By far the weakest folding smartphone I've ever tested. And it gets worse. Straightening it back out for round two, the battery decides it's had enough. Just to explain what happens here, he takes the phone. He bends it backwards with, like, great force. Now, he does it with his hands.

Sure, but, like, hard, sure. The thing he says at the end of the video is, don't put a pixel 10 pro fold open in your back pocket, which is, like, A, I think impossible and B, sure. But he bends the thing back, and it cracks not along the hinge, but along the antenna line, which I thought was really interesting. But then he decides to go do it again. So puts it back together, broken,

and then peels it apart again, and then the battery explodes. And this is because the whole kerfuffle where a lot of people, even he says in the video, this is the first time I've ever had a phone like this explode. The battery goes,

you know, the battery does go nuts. I did this to a phone once with a screwdriver by accident. That's a whole separate story. But then, about all the people who are like defending Google

are like, sure, this is like a completely unreasonable amount of force to put on a phone. And yes, if you break your phone and then break it again,

something bad might happen. the battery. And then there was somebody, I think it, I fixed it, who was like, oh, he probably just didn't discharge the battery properly. We only do this with less than 25% charge in there. And it's like, I don't think that's the point here. No, no, no, no. But so what I wonder from you is like, is this a real scandal? Well, like, you measure scandal by like a drop in sales. And so the title of this YouTube video, and again, I love Jerry Rig everything. Like, I sincerely love this channel. It says, in all caps, my pixel 10 pro fold exploded, caught live on camera. Like, it's being positioned as a scandal

that this is a bad thing that we have revealed. Okay, so my immediate reaction to that is that it's very funny that Google is being hoisted on its own algorithmic bias. Yeah. Sorry, Google, you created the conditions

by which that is the only acceptable YouTube title for this thing. Right? That's very true. This is your fault. Like, sorry. Like, that's what you got. You built this system

and that's, you added these inputs and you got that out. put straightforwardly. So that's funny. Yeah, I think it's a scandal. I think you should, someone's going to bend this phone backwards, and then they might be like,

oh, I can fix it and then do it again. That's pretty, I can just see that, you know? You can just see that happening. Yeah, my phone's broken. What's a little more? No, but like, oh, no, like, can I, can I, you know, like a desperate sort of like, I'll fix it. And then it happens again, and the battery goes,

Like, that's... That's fair. You know? Yeah. I do kind of see what you mean. That's like a very obvious next thing that happens. Especially because, you know, like this... It is weird that it broke on the antenna line.

But the screen is flexible. Right. So like, maybe I can just... Just shove it back together. It'll be fine. Yeah, anyway, the video is very good. And it is... Like, there's just a thing...

Whatever the opposite of ASMR is, that's like people's experience watching him take a, like, an exacto knife to a... glass display. I can't get enough. It's true. I agree with you.

One of the all-time great YouTube bits. It really is. But it is very funny that if Google PR has one ounce of like heartburn about this title, it's like, it's not his fault.

No. You should walk yourself over to the YouTube building and be like, can you make the algorithm not only reward the most outlandish bullshit? Yeah.

It's in all caps. The exclamation point. That's how you win the algorithm. Yeah. It's the only one. All right. You did this to yourselves, is what I'm trying to tell you.

Pretty much. All right, you've been back like 10 minutes, we've already gone way over. Perfect. It's good to have you back. I'm very glad I'm excited. Will Brendan Carr is a dummy? We'll come back next week.

Yeah, with a vengeance. Brendan Carr has continued to be a dummy. We'll catch up on a lot of Brendan Carr next week. I'm very excited about it. In the meantime, that's it for the show. On Sunday, we have our third episode of version history coming out. It's about the Sony Watchman, which is a gadget from 1982. We wanted to see, frankly, if we could make this show about a show about a

gadget. Most of the people in the room and in the audience had never touched before. We had a lot of fun. It's a really good episode. I think you'll enjoy it. On Tuesday, we're going to be talking about the new Xbox ally stuff that spoiler alert sucks. We're going to be talking about that. Plus, Hayden Field is going to come on and talk about whether AI is making us stupid, which I think will be a fun next thing to do after all the stuff we were talking about today. Nelai, to your point about the subscription, we're doing a survey, the verge.com slash survey, where you can go and tell us all of what you want and what you hate and your least favorite things about Nelai and everything.

So go take the survey, we'll put a link in the show notes, but take that to tell us everything so we can keep doing more cool stuff for subscribers. If you would like some outlets to exist, they do not feel the pressure of brands breathing down their neck

or getting approval content, which is all stuff that happens out in the world. Pay us some money. I don't know how to save money. Yeah. We have meetings about headlines and titles for things sometimes and there's always the like, this is the title

YouTube once, but I don't want to do that actually, to be honest, I do think we should do more of that. And so the Vergecast, this one is going to be called David gets owned by Nelai. Sative takedowns. And honestly, the more money you give us, the more I will take David down on this show.

All right. If you want to take me down, you can email Vergecast, theverge.com, call the hotline 866, Verge 11. This show is a production of The Verge and the Vox Media Podcast Network. The show is produced by Eric Gomez, Brandon Kiefer, and Travis Larchuk. We will see you next week. Neelai, it's great to have you back.

You want to say it? Rock and roll.