The new Xbox is not an Xbox

32:56We need to take a break and then we're going to come back and we're going to talk about what AI is doing to our brands. Be right back.

33:18in 10 % of all e-commerce in the US, from household names like Mattel and Gymshark to brands just getting started. And they want to help with everything from sales to design to marketing. Can't code? No problem. Shopify's got slick, ready-made templates that actually look good. If you need help writing product descriptions or touching up photos, they've got you covered there too with their built-in AI tools. And if you hit a wall, their 24-7 support team is right there for you. If you're ready to sell, you're ready for Shopify. Turn your big business idea into reality with Shopify on your side.

33:53Sign up for your$1 per month trial and start selling today at shopify.com slash VergeCast. Go to shopify.com slash VergeCast. That's shopify.com slash VergeCast.

34:11Support for the show comes from Framer. In the olden days, back in the late 1900s, if you wanted a website, you either had to learn how to code or you had to pay a lot of money to hire someone to do it for you. But today, we have the luxury of website builders that can do the hard work for you. And if you use Framer, you get a full-featured, professional design tool completely free. Framer already built the fastest way to publish beautiful, production-ready websites, and it's now redefining how we design for the web. With the recent launch of Design Pages, a free canvas-based design tool, Framer is more than a site builder.

34:47It's a true all-in-one design platform, from social assets to campaign visuals to vectors and icons, all the way to a live site. Framer is where ideas go to live, start to finish. Framer even helps you design more than websites, creates social assets, campaign visuals, icons, and even site resources all in one place. Ready to design, iterate, and publish all in one tool? Start creating for free at framer.com slash design and use code VERGE for a free month of Framer Pro. That's framer.com slash design and use promo code VERGE. Framer.com slash design, promo code VERGE. Rules and restrictions may apply.

35:37Support for this show comes from Wealthfront. When markets feel unpredictable, finding a safe place for your money can be hard. With Wealthfront's cash account, earn a 4 % APY on your uninvested cash from program banks with no minimum balance or account fees. Plus, you get free, instant withdrawals to eligible accounts every day. So your money is always accessible when you need it. No matter your goals, Wealthfront gives you flexibility and security. Right now, open your first cash account with a$500 deposit and get a$50 bonus at Wealthfront.com slash Verge. That's Wealthfront.com slash Verge.

36:14Bonus terms and conditions apply. Cash account offered by Wealthfront Brokerage, LLC. Member FINRA, SPIC, not a bank. Annual percentage yield of deposits as of September 26, 2025 is representative, subject to change, and requires no minimum. funds are swept to program banks where they earn the variable APY.

36:38All right, we're back. Hayden Field is here. Hi, Hayden. Hi, happy to be here. Lots of you on the Vergecast recently. This makes me very happy. It's been a blast. I'll come on anytime. It's just AI shenanigans everywhere you look. This is what we do here now. Every single day. So the sort of big, heady question I want to spend some time talking about, And I think this is a thing we're like as a society going to spend a lot of time talking about is essentially what AI is doing to our brains and our ability to use them. And I think we're at this fascinating moment where like these these tools like ChatGPT and the rest have been around long enough that we're starting to get real sort of useful longitudinal research on like what it means to use them and how we're using them.

37:20And we're past some of the like, you know, thrilling novelty of it all. And so we're starting to get like real actual evidence about how they are changing the way that we work and think and all this stuff. but also we're still so early in this phase that it's like we don't know yet if this is just the same moral panic that happens every time something happens right like rewind to any new anything and everybody thought it was going to destroy our brains right like calculators were going to destroy our brains and the printing press was going to destroy our brains and newspapers were going to destroy our brains and the radio was going to destroy like pick anything you want and it was always going to ruin society and most of them have not um some of them have but uh we don't know yet which one AI is, but I think especially over the last few months, we've gotten a couple of really interesting studies that have tried to dig into this stuff in a pretty meaningful way.

38:07And just before we started recording, I told you one of the main ones I want to talk about is this big study out of MIT that basically was looking at what our brains are doing on ChatGPT and how we think about tasks and how we accomplish tasks when we're given LLMs to use or not. And you just told me that literally just before we started recording, this is like not a bit I'm doing. This is like a thing that happened. I'm very excited about it. Just before we started recording, you were talking to the person who led this study. Is that right? That's right. Yeah. I just shot her an email because I figured, you know, let's get it straight from the horse's mouth and talk about, you know, what really happened here.

38:49and how they tested people and what some media outlets might have gotten wrong. When I was looking at the study, I saw that they had only linked to certain outlets because they said some had like sensationalized their findings. So I just wanted to make sure we really got right, like what they found in this study. So just to really quickly, let me just like briefly explain the study for people who have not read. It's like 206 pages. You should read all of them. They're fascinating. I don't know anything about EEGs. so most of it made no sense but it's very interesting and essentially as i understand it and you should correct me if i'm wrong because i think it sounds like a lot of people are getting this wrong and i'm sure i'm one of them basically they took this group of people divided them into three sections basically there was a group of people that were given access to a large language model there was a group of people that were given access to a search engine and there was a group of people that were given access to neither and they gave them an essay writing task four different times over the course of, I think, several months and basically charted how their brains were working as they were doing those tasks.

39:54And then at the end, on the fourth task, the people who had to use only their brains were given access to an LLM on the fourth try and vice versa. The people who had had LLM access the first three then had to just use their brains for the fourth. And we should get to what they find. But I'm curious what you what you learned from talking to the person that led it. Yeah, they did swap up the groups on the last try, which was interesting. And the thing that I thought was the most interesting is that when they swapped the groups at the end, the same exact results, which we'll talk about in a second, happened.

40:29So it wasn't like if you had been doing something one way, there was a compounding effect or vice versa. It was whenever you did one thing, the same thing happened. And whenever someone else did that thing, that thing happened. And there was no, you know, compounding problems or benefits. It was just when you did it using XYZ type of technology or no type of technology, they saw a pretty similar result. Okay. And I would say, broad strokes, my understanding of the results of this study, and this has been a pretty controversial study. It's a relatively small number of people. It's a day they released this pre peer review, which is is a like peer review is an important part of the academic process.

41:08Like this is an early thing. And I think they put it out basically saying, yes, this is early, but we think it's important to start getting this stuff out there. But basically what I would say, my understanding of the findings, is that the people who had the most access to the LLMs came to rely on them almost completely over the course of doing the task. And that the more they had access to these tools, over and over they came back. And by the third time they were doing these essays with the LLMs, they were essentially just copying and pasting the output of the LLMs. And all the people who use them, this is my favorite detail from this, basically ended up with the same essay.

41:45No matter how you started, no matter what you prompted, they said there were people who started in a language other than English and then translated it back into English. Like a remarkably similar thing appeared at the end. And the funniest thing about this is, by all accounts, they had human readers and AI judges, all of whom agreed that the essays they came up with were actually pretty good. But that was not the point, right? The point wasn't, can you use AI to write a good essay? The point was, what is your brain doing in this process? And they found, essentially, that the people with that were doing less and relying on the AI models more.

42:20And that was increasing as time went on. But then, like you said, once they once they flipped back to having to use their brain, it wasn't like they were permanently ruined forever, I think was the thing that I found. And I would guess not to put words in our researchers mouths, but that was the thing that I think a lot of people picked up on incorrectly. Exactly. And I think that the narrative of this study has been that, like, if you use it, your brain will go to mush and you will never be able to use it again. And I think from reading the study, my understanding is like, that's not the finding.

42:51The finding is that when you have it in front of you, you will you will use your brain differently, not it will destroy your brain in perpetuity. Is that fair? Is that your read? Yeah, the researchers said that they recommend like longer term longitudinal impact studies on the usage. But this was just short term, like seeing what parts of your brain lit up or which parts, which brain networks were being more engaged. And that was it, not kind of charting what's going to happen in the future. Right. It's not that the people came out of this unable to spell their names because they'd used an LLM three times.

43:30It was just that when they had it, it was very clear that they were doing less of the critical thinking part of writing an essay and more of prompting an AI to basically give them the same thing that everybody else was getting. I will say this study, a lot of people have not talked about the search engine part of this study, but maybe the single most damning part of this study to me is the people who are given access to a search engine and basically just wrote the essay that Google decided to give them based on how Google's keywords rank for ad prices. That is like just overwhelmingly you search for something and Google gives you whatever it feels like giving you, whether or not that has any bearing on what is like useful and good.

44:09And so the idea that like using AI is bad for your brain and having access to Google is not, disagree. If anything, you could argue that they are like equally bad for your brain. But I do wonder, like, is it a fair read to come out of a study like this for you as a reader of this and for them as the researchers to say it is bad for your brain to have access to these tools? I don't think they'd say it was bad for your brain because your whole brain was functioning the whole time, no matter which group you were in. It was just the connectivity was really different. So you didn't have the same amount of engagement among all your brain networks.

44:48And also something that she told me on the phone, the lead author, Natalia, was that directionality was impacted. So basically, this is a really simplified version for us because we are not researchers on the brain. but something they were tracking was where data flowed which direction to which direction data flowed within the brain and they saw there was a lot less flow when they were when people were using llm so you know if the information is flowing from the front of your brain to the back of your brain the back to the front that's super oversimplified but that's what she told me that looking at that specifically the directionality there was a lot less flow when people were using LLMs, and that's not good in the moment, at least.

45:33You know, it's not, they put it like this, the cognitive cost of using an LLM. So basically, you know, the brain connectivity did systematically scale down when people were using search engines, and even more so when people were using LLMs. So I think that the main kind of takeaway is that using your brain on its own is the best for your brain probably, at least in the moment, because, you know, those people, that group exhibited the strongest, widest ranging brain network activity. So clearly, you know, their brains were just lighting up. It was a party in there. There were a ton of thoughts happening.

46:12Other people, less so. But also, I think it's important to note this isn't just happening with AI, like you said, David. It's like, you know, I remember when we got GPS systems in our cars for the first time, and I forgot how to drive to like, you know, the grocery store or to one of my friend's houses. I didn't remember how to do things that I used to do if I didn't practice. So it's kind of like that. I mean, I think it does happen with every technology. You know, obviously brain connectivity also scaled down in this study with search engines. So I think it's just about, you know, I used to kind of roll my eyes when my friends would put their phone down and say, oh, like I'm going to try to remember the name of this person instead of looking it up in my phone.

46:51Or I think I'm going I try to remember the name of this movie instead of Googling it. But now I'm like, you know, maybe we should all be doing that because, I mean, it does use a different muscle or whatever part of the brain that you and I don't know how to refer to. It uses that in a different way. And, you know, under-engagement is something they saw in, and weaker neural connectivity was something they saw in the group using LLMs, which I just don't love the sound of. Yeah, that's fair. I just want everyone listening to and watching this, by the way, to just imagine that every two minutes, like a little asterisk appears above our heads that just says, disclaimer, not a brain scientist.

47:28Yes. That's important. But I think like the way you described that makes me think of, actually, I was just reading another study that Microsoft researchers did earlier this year, basically looking at how people doing knowledge work tasks engage in critical thinking. And they define critical thinking in some really interesting ways. But I think like, however you imagine that in your head is a useful enough heuristic for what we're talking about here. And essentially what they found is that if you trust yourself to do something, like if you think you are an expert in a task, AI is actually like productive and useful and helpful in that task, and it will allow you to do it better.

48:07That it is a way of taking something you do well and magnifying it even further. But AI as a replacement, like as a way to fill in something you don't know how to do or don't have confidence in is a huge problem because it a it makes you more confident in this deeply imperfect technology and you just kind of plug in a prompt and let it do its thing and trust whatever it comes out with because you don't know any better and so over and over they're coming in this study basically being like what we need to figure out how to do is a help people gain competence themselves and be plug in AI in a way that is like additive and productive, not a solution so that you don't have to use your own brain.

48:51And I think like the way you describe I was it's so funny you mentioned the like trying to remember a person's name, because this is a thing I go through all the time now. And there's like, people have been talking about this for forever, right? That Google like ruined the bar debate about who was, you know, who was in that movie or like which one of those made more money or whatever like all of these things that just have answers but are fun to talk about anyway doing that on purpose is like stupid but kind of fun and productive anyway that it's like i'm just gonna sit here and try to remember the name of that song even though i could google it and it would simplify everything because like i should probably work on my brain like feels ridiculous but i think is actually borne out in a lot of this stuff that like using your brain on purpose is good.

49:34Shocking, shocking, but true. Yeah. And what you said about AI making some things better is true because even in that MIT Media Lab study, they found that it does help with speed and execution and grunt work and organization of thoughts. And the other thing about that study is that a lot of neuroscientists apparently say the EEGs, which they use to test the brain activity, are kind of a flawed method because they're outside the brain. So it's kind of like they liken it to standing outside a stadium and listening to the crowd's reaction. So you can hear when the crowd roars and, you know, the vibes that are going on, but you can't really, you know, figure out exactly what's going on and, you know, exactly what people are feeling and when and which areas.

50:22So, you know, it's a flawed approach, but, you know, it does tell us something. So I think that's important. But what you said about, you know, feeling confident, I think is so important because it reminds me of learning languages. Like if you are super, you know, playing it safe and taking it easy with learning a language, I feel like you never really, and I think studies back this up, but don't quote me on this, but you never really get fluent because the way that the brain learns a lot of times is when you do feel uncomfortable or frustrated or not like safe in terms of, you know, saying what you want to say.

51:00And then it causes you to build new neural pathways and remember, you know, connections and things you might not have gotten before. That's why immersion programs are so good for building language fluency. It's because you're forced to just figure out how to say something, even if you have to be creative and kind of talk your way around something or, you know, build new connections. You just have to figure it out. And that's why people that get fluent in a language do that stuff. And I think it's kind of important to note here, too. Totally. So what do you think the reaction has been in the AI industry to some of this research?

51:33I mean, this MIT study we talked about on purpose because this like it caused some real like debate and kerfuffle in the AI industry. There were a lot of people who just reflexively like didn't like the study, didn't like the way that it was done, didn't like the product, like whatever. But this question of like, is I, is AI making us stupid is very much like in the ether of the AI industry right now. What are you hearing from folks who are starting to reckon with this question of like, what, what is AI doing to people's brands? Are they even reckoning with it at all? I think a lot of times they're not reckoning with it.

52:08Like when I I reached out to Anthropic and OpenAI before this recording, like a few days ago. I just wanted to see what they would say, and they did not respond at all. So zero response. And I think, you know, a lot of people just don't want to acknowledge it. I do think there's a lot of controversy over the study itself and the results and the small sample size and the EEGs. But I also think that, you know, there's a question why we're not seeing more research starting right now about the long term effects of using these systems. And, you know, there's a lot of questions out there right now about how it does affect your brain, or even if it doesn't affect your brain in a certain way, how it affects how you think or how you act.

52:50You know, we saw a bunch of hugely problematic situations with teens and young people recently, you know, committing suicide after months of confiding in a chatbot or, you know, going through a lot of mental health struggles and having to be hospitalized. So it's like interesting and just, you know, we need more research on this stuff and we need to be looking really hard at it. And a lot of companies don't want to reckon deeply with it because companies want people to use their services and rely on them. I mean, that's how you make money as a tech company. So, you know, up engagement, up the number, the amount of time that someone spends on a platform, figure out how to kind of insert it into every part of their life.

53:38You know, we just saw OpenAI launch Pulse, which is supposed to be like, you know, the way that you start your day every day and a routine with chat GPT. So I think these companies want their products to be embedded in every part of your life. And, you know, that is how they're going to make money. Yeah, I mean, I'm so torn on this, because I think like your Google Maps example from a few minutes ago is such an interesting one, because I spent a lot of time over the years wondering whether it's bad that I don't know how to get anywhere anymore. And like, it's sort of seriously, like, I don't know how to get nearly as many places as I used to.

54:13That's certainly true. And the number of places like my wife will make fun of me because I will Google Maps the three miles to the airport, despite the fact that I've done that drive 650 ,000 times, like, I should be able to do that drive in my sleep. And I put on Google Maps every single time. And I say like, like, oh, I'm checking for traffic, but I'm just like, I'm just not 100 % sure how to do any of that anymore. And I don't, I don't know if that's bad. And I think in a lot of ways, like, I don't know, like you and I probably know how to do less long division than people who grew up before graphing calculators existed.

54:47Is that, is that bad? Like, or have we just learned how to use our brains in a different way? I'm just curious for you as a person who like thinks about and writes about and uses AI stuff all day. Are you thinking about this stuff in sort of that big, heady, like what is this doing to my brain kind of way yet? Yeah, to me, I think about it right now in the way that I think about Google Maps, like you just said. I mean, I think we need more research on the long-term effects, but for now I just think of it as, oh, I can tell I'm not using my brain as much right now, but I don't think there's a long-term effect happening.

55:27And I also do think it's really helpful for, like the MIT Media Lab study said, like grunt work organization. Like I was making a wedding planning spreadsheet the other day and I was like, I don't need to be, you know, looking at five other people's spreadsheets and trying to, you know, combine them myself. I'm just going to have AI do it. So I did. It was great. And then I still critically thought as much as I could critically think about, like, you know, the certain things I needed to add in there that were missing. But yeah, I mean, for me right now, it's more about making sure I don't over-rely on these systems when I'm using them in my personal life.

56:03I turn off memory, for example, because I don't really want a system learning more about me and tailoring its responses to me. I like to treat it as a tool that I ask something and then I leave. You know, it's not, I don't want to be building a relationship with this tool. I know that's controversial. A lot of people do love to do that. But for me personally, I don't want to over rely on these systems. So that's what I do. And I think it's a great hack if you are worried about that. But yeah, I mean, for me, I think that's the number one way I think about it. It's just how I think about these systems rather than what they're doing to my brain.

56:36Because I think, you know what, I'm frying my brain with a bunch of different technology. It's not just AI. I think I'm watching too much Netflix. I'm, you know, texting too much. I am on Instagram too much. I'm on TikTok too much. So yeah, I think like everything in moderation is, is as usual, pretty good advice. Cause I don't know. I think about this as like, there's one way of thinking about this that is like the people who say like, oh, we used to be hunter gatherers and now we've lost that. And what are we doing? And like, when men were men and I just listened to that and I'm like, yeah, but it's better now.

57:09Like I have, I have video games now. Like that's so much better than being a hunter gatherer. Like what are we talking about? this is i can fight lions or i can watch television like i pick television i don't understand what the problem is and then on the flip side you go all the way to the other end of the spectrum and you get all of these tech ceos who are like in five years there's going to be two types of people there's going to be the people who have made use of ai in every single corner of their lives and there's going to be the people who get left behind and it's like what was it that mark zuckerberg said just the other week like there's you're going to be at a cognitive disadvantage if you don't use AI.

57:43Like, boy, do I not think that's fully true. And I don't know, like, part of me is like, are we going to get to the point where actually, there's going to be the like, WALL-E world of people just like sitting in chairs, staring at screens. And then there's just going to be like, you and me running around being like, haha, I still know how to type words with my fingers. And that's going to be, that's going to be our advantage. I just don't know. And it's a really tricky thing, both as a person and as a reporter right now to reason through because it's like, this stuff is everywhere. People are using it.

58:18And like, I have a nephew who just started college, and I'm sure is just in the middle of chat GPT-ing his way through all of his classes. And I'm like, is that a disaster? Or is that just the inevitable future and everybody needs to catch up? Like some of the pushback I've seen to these studies is that actually what happens when we get these technical advances is that we start to teach different things and we start to test different skills and we start to evaluate people based on new things that they need to be able to do in this new world where being able to quickly do long division is not important because you have a thing in your pocket that can do that for you we we need new tasks and we need new things to do and then maybe we'll get a whole bunch of that with ai or we're all just gonna sit here slack jawed talking to clod for the rest of our lives and that feels bad and i just i don't know i'm so stuck on this because i don't know how to reason through any of it because we are at just the very beginning of actually starting to understand this stuff.

59:13Definitely. I think it's interesting because I remember being in college and using Cliff's notes, you know, to summarize. Knew that's what you were going to say. Yeah, exactly. To summarize books that I didn't really have time to read the whole thing of. And, you know, I do think that teens are going to find workarounds for not doing a lot of work in any universe and with any tech that they have available to them. And it's not something that's like, you know, suddenly happening now for the very first time. I do think it can be tough when these companies allow like every student to access these tools for free during exam period.

59:53Which to your point about like, all they actually want is engagement and they'll figure out the rest later. Like what a perfect example. Yeah, I kind of feel like it's, you know, They like to pivot it as, oh, we want people to use learning mode and kind of be helped with some of the grunt work and the organization of their essays, things like that. But what inevitably happens is that people lean on it for exams and all sorts of other stuff and kind of get in early and they become a user leader in life. And it's just kind of cemented early on. So I do think it's a little bit tough, but I also think what you said about testing different things is so real.

1:00:33I mean, you know, now I'm seeing that teachers are starting to, instead of assign a full book, sometimes they'll assign certain chapters or passages and then have everyone talk about them live in class. So that's something I like to see. I mean, obviously, I wish everyone was reading full books. But, you know, at least we're, you know, having people really talk about their what they critically thought live during class and, you know, use their brain that way. So, you know, I think it's hard because it's not black and white. But, you know, at least teachers who are saints are adapting to this new world and trying to get students to use their brains in new ways, you know, no matter what's thrown at them, which is great.

1:01:14whose responsibility do you think it is ultimately to start pushing on this stuff like obviously i think you talked to the the researchers who are doing some really interesting and important work and it's very clear that like a lot of people are paying attention to that work which is great like whatever natalia and that team do next a lot of people will pay attention to and i think that's great like putting all of the onus on like the high schoolers to make good responsible decisions about how they use their brain, I think is like a complete fool's errand of an idea. Teachers have too much to do already.

1:01:47Like, I don't know. I can't help but wonder, like, isn't this partly up to the Sam Altmans and Dario Amadeys of the world to like, if these things are going to be as big and important as they claim, shouldn't, don't they have some responsibility to figure out like, what is the actual job we are supposed to put these things to? And what does it mean if we're just mushing everybody's brains in service of giving them more access to ChatGPT? Yeah, this is kind of why I think it's weird that only a few people are in charge of the implications societally of this. It's interesting that, you know, a lot of the regulation is just starting to kind of percolate right now.

1:02:34And a lot of the past regulation has just been reporters, you know, shining light on stuff or researchers shining light on stuff. So it is interesting, especially because we don't have, as we said before, a lot of information on how this looks long term and how it affects people. You know, I mean, I think on the one hand, you know, these companies are saying, hey, we introduced learning mode. We introduced these certain modes for students where, you know, the chatbot isn't just giving them the answer. It's helping them reason through, you know, a line of thinking. It's kind of using the Socratic method in a way.

1:03:11But how well can it really do that? I haven't tried it. And, you know, I'm sure there's a workaround for not doing that, which students will probably find, you know. On the one hand, it is a student's responsibility in some ways to want to learn because we've all found workarounds when we didn't want to learn something. That's just, you know, what we're going to do. But I I also think the easier it is to not learn something, the more tempting it is to not sometimes if you're like, you know, a student in college or high school. Yeah, not doing the work is so sick. Like, it's so fun to not do the work.

1:03:46And we were actually just talking about this to some folks the other day that, like, one of my great joys as an adult has been going back and reading all the books that I fake read in high school. Same. And learning that actually, like, Ketcher in the Rye is, like, awesome. And there's like a bunch of like, you know, Jane Eyre style books that I like pretended to read in high school and college that now I've gone back and I'm like, oh, these books actually like kind of ripped, dude. I've been doing the same thing like Beloved by Toni Morrison. I've been really I saved them all to put in my bookshelf in the living room, which is my performative bookshelf.

1:04:16And I have like a non-performative, like kind of like dumb bookshelf in my bedroom and all the ones on display in the living room. I'm now working my way through. I love that for you. That's awesome. So one last thing, the way you were talking about these tools makes me think about like, I think the like notebook LMs of the world. I wonder if things like that end up being sort of the right middle ground where it's like you. And maybe this goes back to the stuff that Microsoft was talking about that it's like these things should be additive tools for things that you are already gaining competence in, right?

1:04:50That like maybe the AI's job is to help you find information that you're looking for or pull things into study guides and flashcards and whatever. But like it shouldn't be able to put out a finished product for you. And I think like I know the Notebook LM team has spent a lot of time thinking about like what does it mean to help you do the work versus do the work for you. And I just think that kind of thing like it's really easy for me to come out of these things being like just shut it all off. turn off the internet in high schools and like make good make everybody go back to writing by hand in blue books and like there's part of me that thinks that but most of me thinks that's just like a ignorant way of thinking about what the world actually is now and so it's like maybe what if what is the middle ground does it exist do you think it exists i hope so i hope it does and i think based on history you know and other technologies i think it does i mean and i think what you just said is the middle ground.

1:05:51Technology can help us come to our own conclusions. It can ask smart questions sometimes and have us think and really figure out what our opinion is and then we answer. So I hope that's the line we go down because everyone has a friend that's amazing at asking questions and kind of helps you really figure out how you feel about something when you didn't really know before. And I hope that, you know, in schools and with students, AI can become that. If we do have to use it, let's make it that instead of just, you know, giving you the answer all the time, which I think, you know, to their credit, companies are trying to figure out how to do.

1:06:34But I don't know if it's working yet. I have not used them myself. That would be a good thing for us to test. But yeah, I mean, it's all a mess right now. And I just hope it works out is all I to say that's that's a hopeful message to end on yeah i just keep thinking about the the that part of the mit study that is basically like when we got to the end of this trial with folks the people who had llm access weren't doing any work anymore they were they were essentially doing the the like procedural work of copying and pasting stuff in an order that kind of made sense and that was the amount of actual cognitive interest that they were doing because they had just decided to offload all of it to the machines.

1:07:14And I think anybody who has used Google's AI overviews knows that we're super not there yet. I don't know if we ever will be. I don't know if it's a good thing, but we're super not there yet. And the fact that they couldn't quote essays that they wrote minutes before. So it really didn't sink in, which sucks. Like, that's the thing that I feel like we need to see is even if people use technology to accomplish whatever in schools, you have to be able to explain it and talk about it after so that it sinks in. Maybe that's the answer, you know, because I find that for me, I learn the best when I'm explaining how something works to someone else, like out loud.

1:07:51That's how I learn. And that's why like oftentimes before I'm, you know, going to go on, you know, a TV segment or something, I call a friend and I'm like, let me just explain this to you really quick. And they are like, you know, they work in like marketing or like they're like a hiker and they're like a full-time hiker. And they're like, I'm going to use a different example because that sounds like I made that up, but that was real. And they work in marketing or they're like, you know, just retired. And I'm like, let me tell you exactly how this random thing works that you have no interest in just so that I can know that I fully understand it before I go and explain it to a lot of people.

1:08:26So yeah, I sound like a comedian developing like their type five to go on a late night show. I love this for you. All right. We need to take a break, but I just want to prepare you for this, Hayden. At some point, you are going to come back on the show and you are going to walk us through the technology and AI of your wedding planning process. So just be ready because I have a feeling it is spectacular and insane. And I'm very, very excited to talk about it. I can't wait to tell you all about it. It is crazy. All right, we got to take one more break. And then we're going to come back and answer a question from the forecast hotline.

1:08:55Be right back. Support for this show comes from LinkedIn. When you're a small business owner, your business is on your mind 24-7. So when you're hiring, you need a partner that works just as hard as you do. That hiring partner is LinkedIn Jobs. When you clock out, LinkedIn clocks in. LinkedIn makes it easy to post your job for free, share it with your network, and get qualified candidates that you can manage all in one place. LinkedIn's new features can help you write job descriptions and then quickly get your job in front of the right people with deep candidate insights. Either post your job for free or paid to promote.

1:09:33Promoted jobs get three times more qualified applicants. At the end of the day, the most important thing to your small business is the quality of candidates. And with LinkedIn, you can feel confident that you're getting the best. Based on LinkedIn data, 72 % of SMBs using LinkedIn say that it helps them find high quality candidates. Find out why more than two and a half million small businesses use LinkedIn for hiring today. Find your next great higher on LinkedIn. Post your job for free at linkedin.com slash track. That's linkedin.com slash track to post your job for free. Terms and conditions apply.

1:10:09Support for this show comes from Anthropic. Claude is the AI built for developers who don't settle for good enough. It's not just another chatbot. It's a coding collaborator that works alongside you. Claude can help you reason through problems, explore alternatives, and arrive at cleaner, smarter solutions. And with MCP connectors, you can plug Claude directly into your workflow, GitHub, Jira, Notion, Google Workspace, and more, turning it into your coding command center with full context across your projects. And with Claude Code, you're not just getting an assistant, you're getting an agent built for the frontier of software development.

1:10:44It combines the latest Claude models with tools optimized for coding, so it can handle project-wide changes, manage dependencies and refactor all while respecting your conventions. The result, you write better code faster without ever giving up ownership of the process. Ready to tackle bigger problems? You can sign up for Claude today and get 50 % off Claude Pro, which includes access to Claude code when you use this link, claude.ai slash verge. That's claude.ai slash verge right now for 50 % off your first three months of Claude Pro. That includes access to all of the features mentioned in today's episode.

1:11:21Claude.ai slash Birch. Every day, millions of customers engage with AI agents like me. We resolve queries fast. We work 24-7 and we're helpful, knowledgeable, and empathetic. We're built to be the voice of the brands we serve. Sierra is the platform for building better, more human customer experiences with AI. No hold music, no generic answers, no frustration. Visit sierra.ai to learn more.

1:12:18like sort of a thing I've been thinking a lot about recently. And I just want to, I just want to feel you out on this, but I didn't want to prepare you at all because I just want to get your gut reaction. Let's hear this question. Hello, BirdCast friends. This is not so much a question as a revelation. I was just listening to the episode where you were talking about Chrome OS and Android and the seamless experience. I am a longtime Android user and Windows user. and have had the tablet that disconnects from the dock that's supposed to be your computer. I've gone through all of it. And I have a revelation.

1:12:58For an Android and Windows user, the perfect experience is a two-screen laptop where the VergeBook Duo, which I have, and the Asus ZenBook Duo, which I have and I love, and the keyboard comes off and the kickstand for the two screens take all of that except the top screen can detach and it is your lightweight Android Chrome tablet. Think of practice. Sean, detachable PCs. Oh my God. We're here, we're talking about hybrid devices. I've been thinking about this where we're talking about touchscreen MacBooks. Like how do you do the Chrome? Like detachable PCs are the future of everything. Are they the future of everything?

1:13:47I feel like I'm being transported right back to 2012. Yes, we're just doing this again. I think I was at CES, and there was a giant fountain. Oh, no, it was Computex because I was in Taipei. And there were just all of these giant, there was a giant detachable all-in-one PC sitting on like a stone pillar next to this giant fountain. And next to that, there was a detachable EPC from Asus. And like the screen would come out of the keyboard region. But the dual screen thing, you're right, the dual screen thing. So I have now played with several dual screen devices or devices with a giant stretchy screen that decides to become two screens when you partition it virtually.

1:14:34And there is really something not to having two screens that I don't care about at all, but having a very big vertical screen in front of you that you can partition into any number of windows you want. but reaching for those screens to touch them is kind of weird so if the top one popped off that would be great the thing i can't figure out is how you would create any kind of hinge and detachment mechanism that would make that not fall apart at the slightest touch because even some of the dual screen pcs i've tried to do screen laptops like those hinges are kind of weird the only one I've ever had that felt really good was I think it was the Surface Book that like the very first one and I don't even know if they made more than one of these but it had a pretty solid hinge that you like you had a little eject button and you pulled it out and then it went in and it really like thunked in and felt like a lot and that was the only one I've ever used that I was like this I could just use and not feel like it was going to break every single time I touched it the only one the only one And I met some people who bought that, and they said that over time, some things didn't always connect up properly electronically, and there were some issues with discrete graphics and things like that.

1:15:55But I don't think they said that the hinge failed on them. The hinge was a really great hinge. Really great hinge, really great detachment mechanism. Do you think chip-wise, we're ready for this world? Because is it possible that in 2012, we just didn't have the internals we needed to do this? But now, with some combination of, you know, internal graphics are getting better, external graphics are getting better and cheaper. Like, the idea of basically, I have a tablet that I dock into a gaming PC seems very powerful and compelling. And, like, it might kind of be possible now. Like, put a Qualcomm Snapdragon Elite into the tablet-y bit of it, and it'll still sort of run like a Windows computer, sort of.

1:16:39dock it and give me like full-on gaming PC stuff. Yeah, USB 4. I mean, USB 4 with the bandwidth, especially with USB 4 version 2 coming out with the 80 gigabit per second bandwidth, I believe. Single port. You plug in your tablet. As long as it's got that kind of a connection in it, you can do the external graphics. You can do the data. You can do the multiple monitors. You can do the power all at the same time. All of that would work fine. and the chips are efficient enough that, you know, if they can fit in a handheld, if they can fit in a thick tablet, they can fit in that tablet too. Making the top tablet, if you want that to be like the entire computer, if you want that to be the entire computer, that thin and light, you might need to look at ARM, like David is saying.

1:17:22You might need to be Qualcomm. And in that case, I don't know how much of the USB for external graphics kind of goodness you get. I don't think they've figured out external graphics yet on the Qualcomm stuff. So if it were AMD or Intel plus external, that would be a little bit easier. I think that this detachable tablet, if you don't care very much about how long it lasts when it's away from the rest of the system, like you're taking the sewing off for an hour, you're using it on the couch, then you put it back in so that it can charge and connect. I think you'll have a better time than if you're expecting the tablet to be your eight hour computer.

1:18:01That makes sense. Yeah, it does sort of sound like, as you describe it, I'm thinking, oh, what this actually is is just two computers bolted together. Oh, no. Because you need all the stuff in the one, but then you need all the other more powerful stuff in the other. And now we've just built two computers, and this thing is going to cost like$3 ,000 and not actually solve any of your problems. Which I think is the thing we discovered in 2012. Please, no. What we need is better software for multiple devices, not hardware that shoves them all together. do not please do not give the designers ideas david what if they had all of the processors what if i just put all the possible chips into a device and we just see what happens it could work so but but i think about this from both a like productivity perspective where like i like the idea of you know being able to instead of like the the the ipad magic keyboard to me is the perfect example of this.

1:18:57I feel like Apple has gone way out of its way to have this funky hinge and that's all well and good, but I would rather have something that looks like a MacBook that I can just pull the two pieces apart. That to me is like, that's, that's where something cool and interesting starts to happen. And the same with a lot of these other devices, but I think about this with gaming in particular, that is like, is there, is there a reason that i might want something like external graphics for you know that's worth spending the money on so that i can have this truly like high-end pc thing that i can just rip off the top half of and take to the coffee shop and i think a lot of companies and a lot of users are like scared about of mechanical points of failure like you don't want too many mechanical points of failure on anything because inevitably those will fail their support calls and so on you have to engineer the crap out of it like that microsoft surface book but i wanted i want everything to be modular i love this idea of i saw this crazy thing at cest this last year where it was an intel uh lunar lake lunar lake may panther one of those two a tablet thing a compute module you stick into the laptop frame and or you take it out and stick it into the tablet frame and that's your tablet and they hinge together and there's like mechanical uh motorized things to turn the screen to face you while you're walking around there just a complete concept and i i said well there are like three good ideas in here if anybody wants to actually produce this hardware it would be super cool and just a small handful of nerds like me will love the heck out of it yeah i i both think this is the wrong idea and think i would love it absolutely to pieces which i think is basically where we landed in 2012 like this this thing is so fun and delightful and make sense for almost no one but i'm gonna i have wanted it ever since and this is my hill that i will always die on is modular gadgets which might be impossible and might be useless but they're so fun and i love it is it time to do the modular phone i feel like phones have finally plateau plateaued to the degree that manufacturers just don't know what to do with the hardware anymore the rounded rectangles beside ai at the time when google did project are uh tried to tried to make modular phones a thing, tried to make those cameras modular.

1:21:10There just wasn't enough reason because there was still innovation happening elsewhere. But now that there isn't much hardware innovation elsewhere, can we bring it back? Please bring it back. Just let me attach weird lenses to my phone and you will have my business forever. This is all I ask. All right. Sometime you're going to come back and we're going to talk about Project Dara, but that's probably a version history episode that we'll do at some point. And I'm very excited about that. For now, if someone is making this Chrome Windows thing that actually might seem possible to exist, I would like to hear about it.

1:21:44And until then, Sean, thank you. This was delightful. Thank you. All right, that is it for the Vergecast. Thank you to Sean and Hayden for being here. And thank you, as always, for listening. If you have thoughts or questions or feelings or other mean things you'd like to say about the Xbox Ally and Xbox Ally X, as always, you can call the hotline 866-VERGE-11. You can send us an email at vergecastattheverge.com. Two quick things before we go. One, hopefully you've heard on the show or on the feed by now that we have ad-free podcasts. But if you're a subscriber to The Verge, you can now go to your account settings and get ad-free versions of all of our shows, The Verge Cast, Version History, and Decoder.

1:22:21It's awesome. Thank you to everybody who has been excited about this. This was the number one thing people were asking us for. Super excited to be able to do it. I signed up for the ad-free podcasts. I'm a big fan. Let us know if you're having trouble with it. But it is awesome. Cannot recommend it enough. Thing number two, thanks to everybody who's been reaching out about version history. We're now three episodes deep. We've got five more in this run. We're already working on the next ones. We've gotten tons of really good feedback. So please keep it coming. Things you want to see, things you want us to make episodes about, different ways you want us to do the show.

1:22:53We've gotten really interesting feedback on like the structure and tone and pace of the show. Keep it all coming. We're about to make a bunch more of these and I want to know everything you think about all of these already before we get into it. But for now, we're just going to get out of here. The Verge Cast is a Verge production and part of the Vox Media Podcast Network. This show is produced by Eric Gomez, Brandon Kiefer, and Travis Larchuk. Neal and I will be back on Friday to talk about just all of the news. There's more news. We have Apple reviews. We've got all kinds of stuff going on. AI continues to just AI.

1:23:21I don't even know what to say anymore. We're going to have a lot to talk about. We'll see you then. Rock and roll.

1:23:30And Doug, here we have the Lemu Emu in its natural habitat, helping people customize their car insurance and save hundreds with Liberty Mutual. Fascinating. It's accompanied by his natural ally, Doug. Uh, Lemu? Is that guy with the binoculars watching us? Cut the camera. They see us. Only pay for what you need at LibertyMutual.com. Liberty, Liberty, Liberty, Liberty. Savings vary. Underwritten by Liberty Mutual Insurance Company and affiliates excludes Massachusetts.