Editor-at-large at The Verge who covers consumer technology and product reviews, and writes the Installer newsletter of recommendations.
Co-hosts The Vergecast with Nilay Patel and hosts the Version History podcast. Has also appeared on Decoder with Nilay Patel, Waveform: The MKBHD Podcast and Search Engine.
Pierce explains why Nvidia's financial role matters more to the AI industry right now than its chip design, and why that role went unquestioned while Jensen Huang was in charge.
And all the people who are buying those chips are not going to be immediately, vastly less likely to because Jensen Wong is no longer the CEO.
Nilay Patel
…I'm the person who made those deals. The second you bring in an outsider to be like, let's look under the hood. Let's see if this tiny little toothpick propping up the American economy is any good at all. Then it's a real danger situation. So I think there was only one choice and it was a CFO.
David Pierce
Yeah. And I think what they're going to try to protect, obviously, is something like stability. But I think we talked to Liz on the show earlier this week about companies like Core Weave, which have essentially been able to do big, risky things and take swings and operate under the assumption that if all of this falls apart, NVIDIA will bankroll us because it's useful for NVIDIA for us to continue to exist. And NVIDIA has infinity money that it has been able to just propagate around the industry. And in large part, NVIDIA has done this without any scrutiny or questioning because of Jensen Wong, who has become the person of this industry. And so I think I'm very curious to see how much of the cult of Jensen Wong personality matters here. Like you said, he's the guy who goes on CNBC and says, like, computing is an asset, and people believe it, right? Like, that is not a thing everybody thought was immediately insane. And I think only somebody, to
Nilay Patel
CNBC's credit, they did have me on the next day to just hyperventilate about that sentence for like five years.
David Pierce
It's a bonkers thing to say, but he said it on television and a lot of people went, great idea, Jensen Wong. And there's not a lot of people who could have pulled that off. And I think that to me is the big open question here is like, is NVIDIA NVIDIA because of its chips and technological lead?…
Nilay Patel
I think MetaStack would go up.
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Near the end of an interview about private torrent trackers, Pierce reflects on why they appeal beyond free content: they combine curation and community that streaming and algorithmic discovery have lost.
mean, to your point about the JD Salinger one, right? Like you definitely get to the point with something like this where attention is the enemy. And actually, the more you can just remain out of sight, the easier time you're going to have continuing to be around.
Janko Roettgers
…interesting to write about the motivations and why people are doing this. That was really the puzzing thing for me. Like, why are you putting so much effort into this? Why are you sharing 3,000 files at a time that you don't actually care about just to improve your ratio or something like that?
David Pierce
Do you think, did you come to the conclusion that it's just kind of for the love of the game and the sport of it all that is what people are enjoying?
Janko Roettgers
I think that's one factor. And that's, I think, a big factor for sure. People obviously also pirate stuff and actually watch it and actually consume it. So that's part of it too. And then there's like, you know, people who are into certain types of content and as like trackers for 70s B movies or…
David Pierce
Totally, especially as you get to the bounties and stuff. You can see how that kind of thing would inspire somebody who is very interested in maintaining the metadata of their movie collection to participate in doing that at a much broader scale. That makes total sense to me. Did you get the sense…
Janko Roettgers
I think there's a renewed interest in BitTorrent in general. So when I did a story about 25 years of BitTorrent earlier this year, I talked to some people who are more like involved in sort of the infrastructure for public trackers. And they said, yeah, we're definitely seeing people. Like coming…
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Pierce was discussing a small Kickstarter smartphone priced at $599. He uses its price to argue that memory costs, driven up by AI demand, are killing cheap phones across the industry.
I think there's music too. There is very dedicated trackers, e-learning or something like that. And so there's niches, but I think movies obviously are like one of the biggest draws. Movies and TV shows, yeah.
David Pierce
It's so funny to think about this as not just a repository of movie files, but as a like discovery and curation mechanism. Like in some ways, a lot of what you're describing is the stuff that the internet no longer does a very good job of, which is just showing you good stuff. And so you join this thing and it's like, okay, I'm in one for action movies from the 1980s, because actually a bunch of people who also like action movies from the 1980s are the right people to tell me about action movies from the 1980s. And the community aspect of that is just as important as the movie files. And it's fascinating to me that those two things have come back together. We tried to split them apart, and in fact, they have just come back together in slightly sketchy ways.
Janko Roettgers
Earlier, you made that comparison to Letterboxd because I think there is an interest and a return in more human-curated experiences. Like algorithms work to a degree for media discovery, and we all listen to Spotify on shuffle every now and then, I think. But people are also starting to appreciate…
David Pierce
Totally. It's yeah, this honestly, I worry about how much fun I think I would have in one of these communities. And I'm like, I need to not join one of these trackers because it will take over my life. It'll be a real problem. All right, Janko, thank you for coming on. It's great to see you.…
Janko Roettgers
Sure. Always.
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Anticipating the objection that AI lab revenue is climbing fast, Pierce argues that the revenue is mostly enterprise spending and that companies are starting to ration it. He treats Apple's reported token cap as an early sign of a business-side pullback.
…earnings call was basically like, what we've seen is people come to DoorDash and then they come back to DoorDash. And the idea that, you know, everybody is going to find a new system through which to order DoorDash is just not at all present in the actual reality of being that service, right? Yes.
Nilay Patel
…and data center regulation, I remember being mad about AI, we're about to hit a real slowdown in how everyone thinks and talks about this that is probably good. And I don't mean like a like a pace the frontier or don't kill everyone slowdown. I mean a, hold on, we should try to make things good.
David Pierce
…want to, I want to bring in one more thing to this, which is there are going to be a lot of people who heard everything we just said and are like, sure, all of that is true, but the revenue numbers at these companies are going up really fast. Absolutely correct. Totally agree. Kudos to them for making $12 billion and losing 41. Honestly, good on them. Congratulations. What we're starting to see, there was some news this week from Bloomberg that Apple is lowering its daily cap for engineers from $300 a day to $200 a day in how many AI tokens they can use. And this is a thing we're starting to see. There was this brief period of token maxing. Do you remember this where everybody had leaderboards about who's using the most tokens? And everybody went, oh my God, look at all this. This is where all this money is going. And now they're starting to slow down. So what you're starting to see here, again, is inklings that not only is this stuff not…
Nilay Patel
Well, there is, there is money in it. It's just going in a circle.
David Pierce
Well, right. And it's all NVIDIA's. There's like, there is one individual trillion dollars, and I'm not sure it's real. And if it is, it definitely belongs to NVIDIA.
Nilay Patel
Yeah. Look, I don't think we are predicting a bubble crash here. I think we're, there's just going to be a like a rationalization of are these products good and are they making money? And particularly in consumer, the answer is no and no.
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Pierce was laying out how badly election night went for the AI industry, calling it the worst possible set of outcomes for anyone bullish on data centers. The ballot tally is his evidence that local opposition is now broad and hard to route around.
Right. To like three or four extremely rich people in particular. Yeah. I do think that picture of the CEOs standing on the dais at the inauguration is truly now going to be one of the like iconic photos of the tech industry.
Nilay Patel
Yeah.
David Pierce
And it is, it, I, I am so fascinated by what's about to happen in this next moment because like you already saw Elon Musk has been running around like deleting tweets. Everybody's making the jokes that, you know, Sam Altman is going to like change his pronouns. Like all of this stuff, the run back to woke that might happen as these folks just nakedly go after whatever they think is going to work for them politically. It's just going to be really interesting. But I do think to me, the thing that I looked at coming out of this is: oh, if you are a person in America who is bullish on AI, whether you are trying to build data centers, whether you are trying to put AI in front of people, this is the worst possible set of things that could have happened to you. Like you mentioned the data center ballots. We pulled this out and there were 41 measures in 10 states on the ballot this week on Tuesday. All but one of them went against data centers. Some of them were like, you know, pro-data center ballot measures that failed. Some of them were anti-data center ballot measures that succeeded. All of them, but one…
Nilay Patel
Sure. Democrats and Republicans alike ran on opposing data centers. Democrats in particular ran on opposing big tech and constraining what the AI industry might do. You know, our approach to all this at the verge is always that we do policy, not politics. Like a deep analysis of how the voters in…
David Pierce
Right.
Nilay Patel
And it's like, well, that's just, that's just bribery. Well, and that's
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Pierce had asked whether the midterms were really about data centers, and was told mostly not, though people feel strongly about them. He offers his theory of why they have become such a focal point.
…on the Verge team have been covering this at length over the last year or so, really. And as it comes to this moment, I brought Addie on the show to see how this became a data center election and what happens next, no matter who wins. Let's get into it. Adi Robertson, welcome back to The Vergecast.
Adi Robertson
Hey.
David Pierce
We're recording this on Monday, the day before the midterm elections in the United States. So I have brought you here to give your Nate Silver style predictions for every single race. I'm just kidding. What I want to talk about, and again, I bring this up because it's important to understand that. There are a lot of things we don't know, like who's going to win a lot of these races. I want to prognosticate a little bit at the end, but mostly I want to just dig into a question I've been thinking about and trying to figure out a lot over the last several months, which is how, in fact, data centers became maybe the central issue of, but certainly it seems like the story of this year's midterm elections. So I guess, to what extent do you think these midterms in the United States are actually about data centers? It's obviously a thing we talk about, but we are set up in prime to talk about this a lot. There's a lot going on in the world. Do you think this is in some real way an election about data centers?
Adi Robertson
I mean, I think in some ways we're not going to know that until afterward. But I honestly like, I know that I should not say this as we are the verge, we are the technology site. I think mostly no. Like I think that large part when you look at polls, if you ask people what are their top issues…
David Pierce
Well, and the thing that I struggle with is I think all of those things are really connected, right? Like I was reading over a lot of our coverage and a lot of what folks have been saying. And we have a lot of records from town hall meetings and stuff. And all this stuff is just so tied up in…
Adi Robertson
I think it's that. And I think we have said this on earlier podcasts and people have written about it, but it's also that it's something that you can stop. Like I think that there was a massive tech lash several years ago. And I think the thing people found out is if you're going up against a…
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Discussing Amazon's new 16-inch Echo Show, which can be set to show just a calendar, photos or a smart-home panel, David says what a home display shows while idle is the central design question, and that Amazon had been getting it wrong.
…just have your photos if that's what you want. And you can just have a Fire TV interface if you just want it to be your TV. So yeah, it's definitely the evolution of the smart display. And I feel like it's gone all the way from really, really simple to actually, this is a full-on computer in your
Speaker 2
kitchen.
David Pierce
So, yes, but also this feels to me like Amazon finally figuring out what people actually want in their house, right? Like we give Amazon a lot of crap about its just relentless Amazoning of everything. They're like, oh, any device that you want. Do you want to shop right now? It's like actually not how people experience devices. But this to me is exactly, exactly the way that you should handle a device like this in your house. It's like, what I don't want is just a screen randomly showing me stuff as I attempt to live my life. What I do want is to be able to have access to this kind of ambient computer. And then when it's not doing its job, it does something else for me, right? Like this is the thing. It's so much better to me that I can just say, when this is not in use, I want it to be a photo frame. And it will. Like, just the hoops you have had to jump through to get this to do those basic things for you over time. And the idea with these computers, and I've talked about this with every person who has ever tried to make one of these, and I…
Speaker 1
Yeah, it is. I mean, it is, it feels right because you can customize it now.
Speaker 2
That was one of the
Speaker 1
problems before is that you just got what Amazon gave you. There was a little bit of customization options, but you never could just have what you wanted. And Panos even said this several times at the event. He said, this is what people have been asking for.
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
The hosts were discussing what 'optimized for touch' means on Apple's new touchscreen MacBook Pro. David argues the touchscreen is a minor convenience, and that people who would really use touch already own an iPad.
and that's all very ipady right the idea that if i press and hold on something it sort of expands to fill more space yeah
Speaker 2
…be testing something testing your app and like then you can you have the iphone display right there like in the test and then you're just like tapping on it and doing it right there it's like ah okay also very cool example not the things that people the most people are gonna do but you know useful
David Pierce
…in the same place, the more productive and efficient you can be. And so I think for Apple to look at this and say, okay, this is a thing you're going to do very occasionally in very small amounts is all fine and good. And then I look at like the demo of the person playing the piano on a MacBook and I'm just like, well, A, no one's ever going to do this. This is nothing. This demo is nothing. The person who needs this, and there are people who need it, 100% guaranteed already owns an iPad. This is not a use case that exists for somebody who doesn't already own an iPad. And so that's the part where it's like, maybe this is just Apple feeling like it has to explain the sort of nth degree of what this could be. But it sounds like even the stuff that's being optimized is basically like really tiny visual changes and accept multiple inputs at a time. So you can use 10 fingers instead of one cursor. That might be as far as this ever needs to go in order to be…
Speaker 4
Yeah. And I mean, it's possible. And it would, I think it would make sense then as to like why Apple's so like stringent about like, if you want a pencil, that you buy
Speaker 2
an iPad. You know, it's like, it's, it goes back to like the whole like, you know, buy your mom an iPhone thing, right? The way the way that they try to like separate everything. And yeah, it's to get you to buy more stuff and all that stuff too. But like, I understand why like they want to keep…
David Pierce
That one, that's a really interesting example. I would also put that on the list of doesn't matter, no one's ever going to do it. It's just the idea that somebody who is a sophisticated enough computer user to use spotlight is going to think that the best thing for them to do is take their hands…
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
The hosts were discussing Anthropic's S-1 and its push into sector products like Claude for legal work. Pierce says conversations with people in AI suggest that having the leading model may stop being how companies win.
And that's just a lot of work. And there's a lot of competition for it. Yeah.
David Pierce
And you have to convince all of those people that your tool on top of your model is better than the vast quantity of specific enterprise software they already have, which is hard. Anybody who has ever tried to sell to an IT manager at a company knows that is very, very, very hard work. And they have to do that at such pace and such scale that just seems virtually impossible. All of this has moved faster than anybody expected. So like, who knows? But all of that is like, that is different from we just have to keep building bigger models. So we need to spend a theoretically infinite amount of money on compute. And to that point, actually, there is a sense in a lot of conversations I'm having with AI people and to some extent in this S1 that actually maybe we are going to fairly quickly hit a phase where just having a leading model is not how you win. That having better products or more tuned products or other things that a company like Google is much better set up to do, like vast quantities of customer support is just the thing that's going to be the actual differentiator rather than we have a slightly better model on…
Nilay Patel
very curious about this dynamic. Look, if I was young and ambitious instead of old and committed to the revenue-free career of journalism, I do love it. You would just start a company and you would take some cheap to run, but pretty capable model and you would just run around all the businesses in…
David Pierce
Yeah. And they kind of have to.
Nilay Patel
It's unclear the entire economy needs frontier intelligence.
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
David Pierce summing up the mathematicians' objection to AI labs dumping solved problems on the field: most of the value of research is in how you reach the answer. Without that, he asked, what does the next problem even look like?
…maybe, if it turns out to be true. And all of that other bit is kind of missed as well. And that's, I think, where it's kind of annoying because they've said this most recent drop of sort of 400-odd results. I've been told would take years at least to sort through. Wow. Just on a basic level.
Speaker 1
…really important to me, especially as audios become more integrated into our daily media lives, whether that's music, a movie, and yes, the Vergecast or Virgin History or Decoder. I want it to sound as good as possible. Visit sonos.com and build your system. Sonos, love the sound, feel the system.
David Pierce
…ways, which is there's a thread in a lot of your stories that is people being like, again, look, even if they solved this, even if the thing arrived at the right answer, most of the value of doing this research is in figuring out how to arrive at the right answer, right? We learn new tools, we learn new systems, we learn new methods, and it is the work that then gets us to the next thing. And this assumes that just knowing the answer to the test is enough. And if you give us the answer to the test, you have given us everything that is required. And they keep just being like, no, that's actually not what this is. Because now what's going to happen is there is no next problem for this AI system to be pointed at because we don't know how we got to the answer to this one. And like, are we now betting on the next open AI model to be able to creatively imagine new kinds of problems, which I think is sort of a leap beyond anything that has been able to happen before? But there's this real sense of people being…
Robert Hart
Yeah, I mean, I think that's almost at the heart of the problem here: you have one side who are like, haven't we done a great thing? And it is great.
David Pierce
We answered your question. Yeah.
Robert Hart
But it's almost like, but that wasn't what we were trying to do. That wasn't all of what we were trying to do. And you've kind of like made the rest of it worse, maybe on the way. That's the thing you can't even tell because it's so like some of the papers, I mean, three of those papers that they…
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
In an episode on the history of the overhead projector, the host explains how 3M got it into schools: by bundling transparencies with projectors, and textbooks and lesson plans made for them. He calls it one of the main lessons of a season spent on classroom technology.
…and they've been a business thing. Now they start to really land in the classroom. At this point, for, as far as I can tell, a pretty long time, the chalkboard had been the main character in the classroom. Is that right, Antero? It was this was this was the thing that everybody used everywhere.
Speaker 3
…orient people towards particular spaces and classrooms, right? Like there's a fixed nature of what the quote-unquote front of a classroom is because of the nature of a chalkboard and then the overhead projector, which gave us much more flexibility than other kinds of devices that would come later.
David Pierce
…were also very good at this argument is that. What 3M started doing was bundling transparencies with both projectors so that you could sort of easily start using them and with textbooks. So they would ship textbooks and sort of lesson plans that were made to go on the overhead projector. And I think one of my great learnings from all of the research for this season of Version History has been anything people can do to make teachers' lives a little bit easier turns out to be very powerful and very hard to walk away from because teachers are so overwhelmed, so overloaded, have too much to do all the time. And if you can just find a way to make a teacher's life a little easier, you can almost always worm your way into a classroom for better or for worse. And there are some of both of those in these episodes. But that just seems to be part of this. And 3M understood that and was like, well, we're going to make one little part of your teaching life easier. And it really seemed to help. As all of this is happening, we need to shout out one other very important and very influential use for overhead projectors, which was bowling alleys. There was…
Jake Kastrenakes
I just, I never believed there was anything before the TV. The TV was there for all of history. So
David Pierce
there was. Fun fact. And it was called the Telescore. And basically, the way it worked is it was this large desk, and the desk was also an overhead projector. And you would sit down and you would program your scores. It would then project all the way down the lane, two lanes at a time, the names…
Speaker 2
Support for the show comes from Granola. Gut check: are you absolutely certain you left your most recent meeting knowing exactly what you needed to do? Are you positive you can even read your own notes from that meeting? If not, you might want to try out Granola. Granola is an AI-powered notepad…
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Pierce is telling the story of LeapFrog's LeapPad, an education-only tablet pitched as the cheap, sturdy alternative to the iPad. What killed it, in his account, wasn't a rival kids' product but the hand-me-down iPad.
But their big idea here, and they actually, they sort of pit this thing against the iPad. They're like, you don't want a general purpose, really expensive, really breakable tablet. You want a tablet. Everybody agrees tablets are awesome that's for education. And this kind of works. It costs $100. That ends up being really important. They sell a few million of them. Like, it definitely counts as a hit. But then the most interesting thing happens, which is everybody who bought an iPad buys a new iPad and they get their old iPad to their kid. And that absolutely kills the Leap Pad.
Jennifer Pattison Tuohy
I remember the Leap Pad. That was very popular for my daughter's age group. It
David Pierce
was. It definitely, it had a real moment for sure. It, I think, got lapped by the iPad, which got a bunch of educational software, which started to be used in schools, which got a million immediate cheaper competitors, even as sort of general purpose tablets. And I think LeapPad just never, never…
Andru Marino
Turning a page in a book, I think they want the kids to experience that.
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
The Vergecast hosts were discussing Sam Altman's remark that society should accept some bad things in exchange for AI's benefits. Pierce's complaint was that the industry is fluent about the risks but can't articulate the upside: "It's going to put everybody out of jobs. It might kill all of us. What's the good part?"
do sick math. And that's cool. I like math. But yeah, it's going to monitor your email for class action settlements. I just, I don't know. Again, just as a watcher of this over the last couple of months, I just kept looking out and it's like, nobody has a better idea for AI than it can help you book trips. And everybody has ideas about how it might kill us all. And like, boy, is that a bad trade. If you can give me the trade that is good enough that I will accept a 10% chance that we're all going to die, that's actually a really fun conversation to have. I would really, really like someone to make that argument once. And no one can do it, including Sam Altman. And it tries to be.
Speaker 3
They're going to cure all diseases, but they also might kill you. Even
Speaker 2
say that. Jemisisabis is out here being like, we're going to cure all diseases. Sam Altman doesn't, he's like, I don't know, war, weapons, maybe. I don't know. And I think that to me was just like this, this thing that like, yes, there are negative consequences to technological innovation is just…
Speaker 3
famous memo from Boz at Meta called The Ugly. And he's like, here are all of the bad things that will happen if we connect everyone in the world and we're going to do it anyway. And there's a lot of lore around this letter. You should Google it and read it. But the sort of most important part of…
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
Closing out an episode on One Laptop Per Child's XO-1, the $100 machine that was going to educate the world's children and instead wound down around 2014. Pierce is naming the belief he thinks doomed it: that technology is good by definition, and that handing people computers lets them solve their own problems. Adi Robertson pushes back in the next turn, arguing OLPC's version of that belief at least wasn't a commercial land grab.
…really ask the countries if that's what they want. Like that quote about we don't even have chairs. A lot of the time, just giving people money or like getting people to be able to make their own decisions, but with more resources is better than we're shoving our techno-futurism down your throat.
Adi Robertson
…commercial purposes, even if it turned out a bunch of people just wanted to access YouTube on them. So so it goes. That is true. But like, I think this is just in a lot of ways a far more value, less pernicious product than a lot of the companies that we're talking about would end up being. That's
David Pierce
so true. I think the middle of those two things that I think I have found myself in is that sort of inherent belief that technology is universally, unilaterally by definition, a good thing. And there was just this sense that if we just give them all computers, they'll solve all their own problems for them. And I think that is, that's just the part of like the that particular ethos of Silicon Valley, particularly at that time. And frankly, like we're going through this right now with AI too. Like there, there's a set of people who believe that we are one, everyone on earth is one chat GPT login away from being the best versions of themselves. And it's like, buddies, like that's not how this works. And it wasn't how it worked then either, but there was a real earnest belief that it was. And I think. I got to say,
Adi Robertson
I don't know. I think at least the belief that the belief wasn't at the starting point, we are going to give a giant company control over your life, which is basically what Chat GPT is. Like, it cannot be overstated that while you can think it's really dumb, the idea was we're going to teach kids…
David Pierce
Yeah, I think that's fair. But I think you're right. And I hadn't really thought about that until you say it. But there is really nothing evil inside of this little device. Like, I think, I think Nicholas Negroponte was out over his skis in incredible ways. I think he both sold and doomed the…
Adi Robertson
Yeah, I think kind of the worst thing you can say about this is that they failed or that they were in some ways fundamentally misguided about education is an argument that a lot of people made. But I think that, yeah, I feel like this was not malicious. Yeah,
Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
