Is Intel Cooked? + What's Your P(Dyson Sphere)? + Hard Fork Gift Guide

The sudden departure of Intel’s C. E. O. reignites concerns about U. S. dependence on foreign manufacturing.

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Speaker 00

  Support for this podcast comes from Dell Technologies. It's time to put AI to work. Is your business ready? The challenge isn't getting started with AI. It's scaling beyond pilots while maintaining control of your data, governance, infrastructure, and costs. Dell AI Factory provides the integrated foundation you need to scale AI securely and efficiently with one trusted partner, Dell, AI on your terms. From idea to impact. Learn more at dell.com/yourwaytoAI. That's dell.com/yourwaytoAI.

• 0:34Kevin Roose?

  Well, Casey, we got so much feedback on our top 100 iconic technologies list from last week. I don't think we've ever gotten more email about an episode.

• 0:44Casey Newton?

  No, we haven't. And thanks to everybody who wrote in. And I'll say it, you really did find the actual omissions on the 100 iconic technologies. And if we did it again today, I think we would probably make between 12 and 15 swaps.

• 1:00Kevin Roose?

  What were the most frequent suggestions we got of things that we left off the list?

• 1:04Casey Newton?

  Most frequent suggestions were language, glass, lasers.

• 1:10Kevin Roose?

  Lasers, so many lasers.

• 1:12Casey Newton?

  Yep.

• 1:12Kevin Roose?

  A lot of laser fans out there.

• 1:14Casey Newton?

  Wikipedia, the steam engine. And one person said teledildonics. And I thought, I don't know if that really belongs on the list with the rest. But.

• 1:22Kevin Roose?

  I don't even know what that means, but I don't think I want to know what it is.

• 1:26Casey Newton?

  Yeah, you don't want to know about that. You're having a great day. But yeah, so we did definitely miss some. And I guess we'll just consider all of those honorable mentions.

• 1:36Kevin Roose?

  Yeah, it's wild. People love a list. They love to argue. And I think we should do that again sometime.

• 1:43Casey Newton?

  I would love to.

• 1:49Kevin Roose?

  I'm Kevin Roos, a tech columnist at The New York Times.

• 1:51Casey Newton?

  I'm Casey Noon from Platformer.

• 1:53Kevin Roose?

  And this is Hard Fork.

• 1:54Casey Newton?

  This week, what went wrong at Intel? The Times' Don Clark joins us to tell us why Pat Delsinger was ousted and what it means for our chip war with China. Then Kevin and I report back from an AI conference where the subject on everyone's mind was existential doom. And finally, it's the Hard Fork Gift Guide. What'd you get me this year, Kevin?

• 2:14Kevin Roose?

  Lump of coal.

• 2:15Casey Newton?

  Bah. Bah, bye.

• 2:24Kevin Roose?

  Well, Casey, the big news intact this week was a shakeup in the chips industry.

• 2:29Casey Newton?

  That's right, Kevin. It's time for us to get some intel on Intel.

• 2:33Kevin Roose?

  So on Sunday, Intel CEO Pat Gelsinger abruptly left the company after being told by the board that he could either resign or be fired. This was a surprise to a lot of people in the industry. Gelsinger had been trying to remake Intel and sort of turn it around. It's been obviously struggling for the past couple of years. And he didn't get as much time as a lot of people were expecting to see his vision through. And Casey, why do we care about Intel and why do we care about chips?

• 3:01Casey Newton?

  Well, look, Kevin, you cannot tell the story of Silicon Valley without Intel. It was once one of the most dominant tech companies in the world, but it has had a really hard time keeping up over the past couple of decades. First, with the rise of mobile phones and second, with the rise of AI. And so we're interested in sort of how a company that once was so successful seemed to have missed the plot. And then beyond that, there are just big national security implications, right? Intel is a big part of the US's plan to ensure that we will still have chips in the event of a war or some sort of other growing escalation with China. And so because it's the centerpiece of our domestic chip industry, what happens to Intel could matter to all of us.

• 3:46Kevin Roose?

  Yeah, so we wanted to get a clear sense of what was actually happening, what the story was really behind this abrupt shakeup at Intel, and just sort of get a picture of where this whole chips industry is going. So to take a deeper look at the state of the chips industry and what happened at Intel, we've invited on my favorite chips reporter, Don Clark. Don is a freelancer for The New York Times. He is considered by many to be the dean of chips reporting. He's been reporting for The Times and before that for The Wall Street Journal on chips and semiconductors for more than 30 years. He has forgotten more about semiconductors than probably either of us will ever learn. And he is just a person that I'm really excited to talk about. I've been looking for a way to get him on the show for a long time.

• 4:30Casey Newton?

  Let's bring him in.

• 4:38Kevin Roose?

  Don Clark, welcome to Hard Fork.

• 4:39Speaker 03

  Thank you.

• 4:40Kevin Roose?

  So, Don, you've been covering semiconductor for something like 30 years, which is also how long I think it's going to take me to convince Casey that semiconductors are important and interesting. Every time I propose a chips-related segment, he tries to shoot it down because he thinks it's boring. But you are here today to dispel the myth, once and for all, that chips are boring.

• 5:02Speaker 03

  I think that's a, you know, that's a pretty good, I think we're on the same team here.

• 5:06Kevin Roose?

  Yeah.

• 5:07Speaker 03

  You know, the thing about it is, I think people started thinking about it like the steel industry, you know, dull but important.

• 5:13Kevin Roose?

  Yep.

• 5:14Speaker 03

  But the thing about it is, is the chip guys know where everything's going two years before the software guys do. And so knowing chips is a really good thing to know about where the future is headed.

• 5:24Kevin Roose?

  Yeah. So, we want to take a pretty broad look at the state of the chips industry today, but I think we should start with the basics of what happened at Intel over the weekend. So, Don, tell us about that.

• 5:35Speaker 03

  So, Pat Gelsinger came back to Intel in 2021 after 11 years away with a really broad mandate to turn the company around. And basically, after three and a half years, it just isn't working. And the board of directors basically gave him a choice to resign or be fired.

• 5:52Kevin Roose?

  And what was the big problem that Intel has had that they needed to bring someone like Pat back in?

• 5:58Speaker 03

  Well, it's not really one problem, it's a couple of problems. But I think fundamentally, is Intel, it for years, for decades, was the best manufacturer. Its transistors on its chip were the best. It shrunk the transistors and got to the next technology advance faster than everybody else. And this basically had, you know, kind of totally ended by the end of the 2000s. I guess around 2017, you could say, basically, it was done. This was under some mistakes that his predecessors had made. And so, you know, his immediate predecessor, Bob Swan, was a total financial guy, you know, do what Wall Street wants, which was stock buybacks and stuff like that. Pat came in and convinced the board to do a totally different agenda, double down on

• 6:42Speaker 03

  manufacturing, you know, get back to the process. Lead against Taiwan Semiconductor Manufacturing Company over in Taiwan. But he basically took on too much.

• 6:52Casey Newton?

  Yeah.

• 6:53Speaker 03

  He also tried to make, you know, rebuild the US as the manufacturing, chip manufacturing powerhouse. And basically, while all this was happening, Intel's products just weren't really selling like they should. So he'd put all this money into manufacturing and the board in its statement about him basically said, you know, products are really what's important. That's what we have to get back to.

• 7:14Casey Newton?

  Yeah, I want to get a little bit more Intel history here. I'm going to tell you sort of what I feel like I've read about Intel and then you can tell me maybe what's right and what's wrong. My understanding is during the desktop computing era, Intel is on top of the world. They're flying high. As you say, those pennyium chips, everybody's got them. They're doing TV ads. There's a jingle that everyone remembers. And then along comes the mobile phone and something happens where Intel just does not, I guess, think that mobile phones are going to be that big of a deal and kind of misses the boat on it. Do I have at least that part right?

• 7:48Speaker 03

  Yes, but the hindsight is of course 2020. In 2007, nobody knew the iPhone was going to change the world. There were other smartphones on the market. Apple was just getting into a market where other people were in. I remember the day the iPhone was launched and they also launched the Apple TV. And we didn't know as news editors which was more important, which just seems hilarious in retrospect. But we kind of double-barreled the lead on that story. But basically, the problem for Intel is that anything it tried to do new, they had these incredible gross profit margins, you know, not only in PCs, they moved their franchise into servers. So basically, they were serving the entire web. Every time you went to a webpage, Intel was driving that transaction.

• 8:29Speaker 03

  So basically, Paul Odalini and Apple was looking for chips and Steve Jobs insisted on a low price. And Paul just said, you know, we're just not going to make any money at that price. In retrospect, that was a huge mistake because in semiconductors, volume is what matters. You know, the more volume you have in the factory, the more you learn about how to make these chips profitably. And the guy With the most volume gets the most information about how to improve the manufacturing process. So basically, once that happened, TSMC, which was making the ARM chips that people used, just over time, just got better and better and better. And then the die was cast.

• 9:11Casey Newton?

  So missing mobile is sort of big mistake number one. And then along comes artificial intelligence and a new kind of chip starts to get super valuable. And that is the GPU, which of course is made by Nvidia. So is there a story there about how having missed mobile, Intel then misses AI as well?

• 9:31Speaker 03

  Yes, there's quite a big story. The thing about GPUs, it helps to keep in mind is is they are a parallel processing machine. They are designed to paint thousands of pixels on your screen at the same time. So basically, the old Intel model of calculations that did one thing, then another than other, that was supplanted by jobs that could do everything in parallel. So around 2006, Intel actually considered buying Nvidia because they weren't doing well in these discrete graphics chips. But instead, actually was Pat Gelsinger, who led a project called Larrabee, that was basically Intel's attempt to counter the GPU threat. The trouble was what they did was they used their own x86 computing technology.

• 10:14Speaker 03

  So they might have, say, 50 processors on the chip where Nvidia had like 400. So Intels were a lot easier to work with. Everybody knew how to program and that was their that was their strategy. They thought they had this huge software advantage. But Larrabee was a total flop. Meanwhile, Jensen Huang, who is the chief executive of Nvidia, realized that we can make this a general purpose computing engine. So he first started doing the software development that would help people program GPUs for other tasks than graphics. So he started this thing called CUDA, which was a basic software framework that would allow people to use GPUs for other things. So about 2012, These AI guys realized they could do machine learning on these chips and get really incredible results.

• 10:59Speaker 03

  And once that happened about 2012, Jensen turned that whole company around and just totally doubled down on this technology, knowing he had to become an AI company.

• 11:09Kevin Roose?

  So there's a lot of worry right now among people who think about the geopolitics of AI about the fact that the vast majority of the chips that are important in building not only AI but other technologies are manufactured in Taiwan by TSMC. That is seen as a big risk if, for example, mainland China were to invade Taiwan at any point, that we would sort of lose access to the company that makes the vast majority of the chips that go into all the AI systems that we depend on today. Intel has been trying to do some things domestically to move production, and I understand they actually just got a big grant from the federal government under the Chips Act to start building more chips in America.

• 11:53Kevin Roose?

  Why didn't that worry about offshore TSMC chip production result in a bigger windfall for Intel or help to make Pat's job a little bit easier?

• 12:02Speaker 03

  Well, it's interesting because Pat, during the COVID crisis, was really exploiting that fear you're talking about. I mean, everybody was. I mean, you couldn't, your $50,000 Ford truck couldn't be shipped because of a 50 cent part. People's mentality started to change, and this is part of the reason why TSMC was encouraged strongly, starting with the Trump administration, to build a big factory in the US. And they now have built this factory in Arizona. So the question you ask about why didn't this produce a bigger bump for Intel gets back to its attempt to regain leadership in the process node technology. They are still well behind TSMC.

• 12:45Speaker 03

  And I think one of the things that forced Pat Gelsinger out is that they were not starting to get customers for the next generation technology called 18A. In Gelsinger's mind, this is where we reassert parity with the process lead. But he couldn't claim any customers for it. And I have a feeling the board just realized, you know, we're not going to get this revenue that we were assuming. I've heard Gelsinger say the very same thing. We were surprised we didn't get more of a sales bump over the paranoia about offshore production sources.

• 13:22Kevin Roose?

  Hmm. And why has it been, you know, I know, I know you said Intel at one point considered buying Nvidia, decided against it, but Nvidia has been such a massive success story over the past several years. One of the biggest companies in the world now, just phenomenal growth and sales and margins on their chips. You would think that such a fantastic growth curve would attract a lot of competition. And actually, you had a story this week about how many companies are now starting to try to compete with Nvidia. But all of those efforts are fairly early. They are still considered quite dominant in especially the AI space. Why has it been so hard for companies like Intel to compete with Nvidia?

• 14:06Speaker 03

  I think the main reason is that they got this head start. You know, they started building the software part really in 2006, and it takes a long time. So basically, once they got their GPUs out, every new AI model is built on Nvidia. So the best chip maker, you could have the best chip in the world, but you're still going to have that lag. Researchers are going to have developed that new cool thing first on the Nvidia technology. So you're always a bit behind. So that's the key reason.

• 14:36Kevin Roose?

  And by software, you mean this CUDA sort of software that sits on top of the chips and is sort of how you program the chips?

• 14:42Speaker 03

  There's a bunch of layers. CUDA is like the most basic. And then there's things called like that PyTorch, they have these frameworks on top of that. And on top of that, people have done all these special purpose things. And Nvidia has become a huge software player. I mean, they, they claim that they've made like, like 400 domain specific AI things for, you know, say if you were a physicist or a, say a medical doctor wanting to get into AI, you know, Nvidia's got the package for you to get into it. So it's become a huge software company.

• 15:12Casey Newton?

  So, you know, as you say, hindsight is 2020 here, but I'm still curious about the, the run up to Pat Gelsinger being hired to try to turn this company around. We know that they missed mobile. We know that they missed AI, but you know, eventually they realized what was happening. So when you think about why they had to bring in someone to turn the Around how much of it is that they just sort of missed those two transitions? And how much of it is, well, they understood what was happening, but they failed to execute or sort of couldn't catch up in time?

• 15:44Speaker 03

  It's a combination of things and a lot of the problem definitely precedes Pat. So for example, we did a story about how they lost out on AI. And besides the stuff we've talked about about Larrabee and that ancient history around 2016 or 2017, I forget, they bought a company called Nirvana Systems, which had a pretty credible startup chip, but they totally frittered away the lead. Once the guy got there, he this guy, Naveen Rao is now at Databricks. You know, he thought he was going to have this great engine to make his chips and everything kind of went to hell. And then about three years later, they bought this other company called Habana to take their AI accelerator business forward. And that basically it's it gave them another two year lag.

• 16:26Speaker 03

  So when Pat came in, they were basically backing about three or four different horses in the AI race. They had GPUs, they had special purpose AI chips. They were improving their traditional chips to get into AI. And basically, the all these things, they kind of frittered and didn't really go anywhere.

• 16:44Casey Newton?

  We've talked about it in bits, but I'm hoping we can sort of maybe get it all in one answer of when Intel hired Pat Gelsinger, what was Pat's plan to get the company back on track?

• 16:55Speaker 03

  You know, Pat made the board totally. Every single board member had to agree to his strategy, which was doubling down on manufacturing and making a really expensive bet on regaining the process technology lead. And, you know, his thing was to make five jumps in process technologies in four years. Normally each jump takes two years. So that's a massive compression effort. And, you know, they seem to have have gone through the phases. The question is, here we are coming next year to sort of the end of that crusade. And what customers do they have? And, you know, what's going to be the impact?

• 17:31Casey Newton?

  So his predecessor just kind of wants to outsource the manufacturing and Pat Gelsinger says, no, that's not what we do at Intel. At Intel, we're going to make our own stuff. We're going to become engineering leaders again. We're going to sort of have this big moonshot product where our chips are getting better and better faster than anyone has ever done it. He comes in in 2021 and what less than three, four years later, he's gone. So when you heard the news this week, Don, were you surprised? Or do you feel like the writing has been on the wall?

• 18:00Speaker 03

  Well, I feel like we knew he was running out of time, but we didn't think he'd quite run out. For one thing is, what would they do? There's no clear successor to Pat Gelsinger. I mean, who would take on this job? He's almost like the only person in the world who would dare to run this big company with its two major halves, chips and manufacturing. And so, you know, I think we thought if it didn't show us some more fruit, he would have to leave. But we were surprised by the timing. And the way that they did it, you know, they basically didn't announce another CEO. Historically, Intel was quite thoughtful in its succession planning. You knew a long time in advance who the next guy was going to be.

• 18:44Speaker 03

  So it didn't seem like the kind of thing Intel would do. So we actually don't know for sure what the straw that broke the camel's back was, but I think there was one.

• 18:52Kevin Roose?

  I mean, Don, you've been covering Intel for longer than almost anyone on Earth.

• 18:57Speaker 03

  Other planets, too, I think.

• 18:59Kevin Roose?

  Is Intel cooked? Like, is there any coming back for this company in any scenario you can imagine?

• 19:05Speaker 03

  Well, I think that, as Pat always said, there's just incredible demand for chips out there. And they're still the number one seller of chips for personal computers. They're still the number one seller of chips that run the garden variety servers. So they have cash flow, they have chip designers. So one of the obvious things is, okay, you can split the company. So Pat basically laid the groundwork for this. He did various things to make sort of a Chinese wall between the manufacturing and its design. See, one of the big fears is if you're a potential customer for Intel as a manufacturer, aren't they going to tell, design things about my new chips? You know, Pat, one of his lines was, yes, we could separate them fully as independent

• 19:49Speaker 03

  companies, but not yet because our factory business needs all the chips designed by the design company to fill the factory, at least for a time. So basically, it's very likely that Intel will face more pressure to formally split the companies into two. And I think one of the things will be interesting to see is, you know, in this push to products, as the board said in replacing Gelsinger, if they start saying, oh, no, this chip we said next year was going to come out on our process, no, we're actually going to make that at TSMC. I think that kind of thing might happen.

• 20:24Casey Newton?

  And when you say, this confuses me. So when you say products in this sense, products is a chip that Intel designs and manufactures?

• 20:33Speaker 03

  Historically, that's all that they did. Every Intel chip was manufactured by Intel. And that's, that's an old school model that most companies had moved away from. But some chips Intel designs are made at TSMC because it had the advanced production process that it needed. You know, Pat was betting that we can bring almost all these chips inside Intel's factory network, which will, once again, have the best manufacturing technology. And right now, it seems like it's questionable whether Intel does have the best manufacturing technology, or will next year. They still think they will, but you know, there's a lot of headwinds. So that's kind of the scenario I see is, you know, you know, you might end

• 21:16Speaker 03

  up with a design company with much more freedom to pick who makes its chip.

• 21:19Casey Newton?

  Right. About how long ago did they start making chips for other people?

• 21:22Speaker 03

  Well, that was really part, I mean, they've tried this, this is called being a foundry. They've tried this for quite a while, but Gelsinger sort of bet the farm on it.

• 21:32Casey Newton?

  He went foundry mode.

• 21:33Speaker 03

  He went all foundry on us. And, and one of the reasons he did that was because they would never get money from the government as just a maker of their own chips. That would be too, you know, that would be too self-serving. If they become a foundry, though, in the US, you know, they would be serving all kinds of companies. And one thing I should bring up is that while now TSMC and Samsung are making factories here, Intel is the only company that does its research for advanced production processes in the, in the US. So that's kind of a strategic thing. If you're, if you're thinking that, you know, America's, you know, say, military and intelligence might is related partly to advances in manufacturing,

• 22:17Speaker 03

  you want some of that research done in the US. And so that was one of Intel's differentiators in this whole chip sack process was, hey, we're the only one that is doing the research here.

• 22:28Casey Newton?

  Here's what I want to know. You know, Kevin and I are always looking for new business ideas. So if he and I wanted to get together and start a new semiconductor business, and that actually could happen because he truly never stops talking about semiconductors, are we better off building our own semiconductors that we sell to people who need our very good semiconductors? Or should we just be go foundry mode and create a foundry and let other and make chips for other people? Like, which of those is a better business in 2024?

• 22:57Speaker 03

  Well, the trouble was, uh, business B, as you've described it, cost billions and billions and billions of dollars. So.

• 23:02Kevin Roose?

  Well, our podcast is very successful, though.

• 23:04Casey Newton?

  Yeah, it's pretty successful.

• 23:06Speaker 03

  Yes, I, I can see that. I can sense. The, um, but, but it also takes incredible expertise. And, um, you know, the thing about it is it's, I, I can, I consider it like baking. You know, it's like there's ingredients and then there's know-how. And, you know, somebody that didn't know how to do this already essentially could not do it. I mean, it would just be impossible. Even designing chips is a huge art unto itself, but that's also expensive. I mean, we're talking like a couple hundred million to get your first chip. And nobody buys a company for the wonderfulness of its first chip. They buy it for the road map of all the chips that are going to come after that.

• 23:48Speaker 03

  So you're looking at billions of, of money to get, to get just as a chip design company.

• 23:54Casey Newton?

  So it would be easier to just make other people's chips?

• 23:57Speaker 03

  Because, well, it'd be technologically easier, but financially almost impossible.

• 24:02Casey Newton?

  Got it. Okay. Well, so my conclusion from all this, Kevin, is that Intel is screwed.

• 24:09Kevin Roose?

  Yeah, a thing that our listeners might be wondering right now, who are not semiconductor obsessives like you and I, is how does this affect us, right? Because in, in 2020 and 2021, we had this semiconductor shortage. People couldn't get their, uh, you know, their, their cars with all the semiconductors inside. That was a very obvious consumer consequence of something happening in the semiconductor market. How will people like our listeners feel the impact of any of this stuff happening at Intel? Or any of this increased competition in, uh, in the chips world? Are there other ways you think the drama, uh, in the chips market right now might ripple out into mainstream consciousness?

• 24:49Speaker 03

  Well, you have to say for starters, barring of some calamity, you know, the COVID thing was a one-off thing that led to the chips crisis. So if things stay the same, we're definitely going to see more competition in the AI chip world. And I would think that would lead to certainly more chatbots. More companies could afford to set up a chatbot, perhaps. So that's on the positive side. On the Intel side, I think, I'm not sure most consumers will feel much impact. They will keep selling chips for personal computers. But I think the other shoe we're talking about, more competition, that could definitely bring benefits.

• 25:31Casey Newton?

  Well, we don't know what's going to happen with the story of Intel, Kevin, but we do know that whatever happens, Don Clark will be chipping away at it.

• 25:39Kevin Roose?

  Yeah, I think you should throw your name into the hat for the CEO gig at Intel. I think you'd be good at it.

• 25:45Casey Newton?

  He's got my vote.

• 25:45Speaker 03

  Yeah, thank you.

• 25:46Casey Newton?

  All right, thanks. Thank you for coming in today.

• 25:50Speaker 03

  Sure, it was really fun.

• 25:53Casey Newton?

  When we come back, our field trip to an epicenter of AI doom.

• 26:08Speaker 00

  Support for this podcast comes from Dell Technologies. It's time to put AI to work. Is your business ready? The challenge isn't getting started with AI. It's scaling beyond pilots while maintaining control of your data, governance, infrastructure, and costs. Dell AI Factory provides the integrated foundation you need to scale AI securely and efficiently with one trusted partner. Dell, AI on your terms. From idea to impact. Learn more at dell.com/yourwaytoAI. That's dell.com/yourwaytoAI.

• 26:42Speaker 04

  AI spend is growing fast, but most businesses still can't see exactly where that money is going. Ramp gives businesses visibility and control across AI providers with real-time guardrails by team and model, helping ensure spend goes to the highest ROI areas. Stop guessing where your AI spend is going. Get started with Ramp. And if you're new to Ramp, you can get $250 off at ramp.com/hardfork. Go to ramp.com/hardfork. Ramp.com/hardfork.

• 27:13Speaker 05

  Ripple has been powering modern finance for a decade. It's not our first trip around the blockchain, and it won't be our last. We're leading the charge to bridge the gap between existing financial systems and the future of finance. You can have it all: legacy, stability, and digital innovation. Systems that work with your other systems. Because with the right partner, the question isn't crypto or traditional finance. The question is simply, why choose? A new standard of modern finance. It's happening with Ripple. For informational purposes only. Availability of products can vary. May be provided by different Ripple entities.

• 27:48Kevin Roose?

  Well, Casey, you and I recently went to a very interesting event together.

• 27:52Casey Newton?

  We did. We did. Normally, we hang out mostly in the studio, but this was a case where we went out into the real world and we met people and we did a little bit of sort of interviewing and what they used to call in journalism school, shoe leather journalism, Kevin.

• 28:05Kevin Roose?

  Yes, but nobody wears leather shoes anymore. So now it's shoe rubber journalism or something.

• 28:10Casey Newton?

  Yeah.

• 28:11Kevin Roose?

  So this was an AI conference in Berkeley, California called The Curve. And it was a very interesting event. It was put on by a group of AI researchers and enthusiasts. And the way it was sort of announced is that this was going to be a place for people from all sides of the AI spectrum to get together and put aside their differences and find some common ground. Whether you are an AI doomer who believes that AI is going to kill us all, or an accelerationist who believes that we should be going faster. The idea was to kind of bring together these warring tribes of AI and kind of hash out some differences and try to find some common ground. And I was very curious to know sort

• 28:54Kevin Roose?

  of your initial impressions, because I had been to this venue before. It's a venue called Lighthaven. It's a former hotel that has been kind of converted into an event space by a group of rationalists. But you had not been there before, and this was your first time at one of these events. So what did you make of it?

• 29:10Casey Newton?

  I mean, my main takeaway, Kevin, was just that I feel like the rest of the world does not know how deeply that many of the people working to build AI believe that we are on the precipice of a profound transformation. And there's no way to say that without sounding like you're just sort of contributing to the hype. But despite everything that I just said, this was not a hypey conference. This was a bunch of people who took for granted the idea that, you know, a superintelligence is almost here and we're reckoning with the implications of that. And so for me, as somebody who had never spent that much concentrated time with people who had that set of beliefs, it was really eye opening.

• 29:51Kevin Roose?

  Yeah, I mean, this is the kind of place where like you walk in and on the wall is a big poster with some timelines on it. And the timelines go from, you know, present day to the year 2280. And on these timelines are various scenarios, such as how long until 95% of current human remote work can be done more cheaply by an AI? How long until the first year with more than 10% US economic growth because of AI? How long until medical advancements are made until that allow humans to live past 150? And the one that really caught my attention is how long until there is a Dyson sphere

• 30:36Kevin Roose?

  around the sun? Now, Casey, do you know what a Dyson sphere is?

• 30:38Casey Newton?

  No, I didn't. And everyone at the conference made fun of me for not knowing what a Dyson sphere is. So what is it? And honestly, when did you find out what a Dyson sphere is?

• 30:47Kevin Roose?

  I have known what a Dyson sphere is for many years. It's like a staple of science fiction and of kind of these like thought experiments about what happens after the singularity. But basically, this is the idea that you could kind of surround the sun with a bunch of solar cells and just capture all the energy and then we would have no more energy needs.

• 31:05Casey Newton?

  Well, you know, and my first thought, Kevin, is how do you stop the Dyson Sphere from being burned up by the sun?

• 31:11Kevin Roose?

  Great question. We haven't gotten there yet. But people do think that this is one of the sort of far future scenarios that we could see. And so what I found remarkable about this particular poster with all these timelines on it was that you had to like, put a pin in the timeline for when you think this stuff would happen. And even for the Dyson Sphere, which was sort of the most far out futuristic thing, I was surprised by how many predictions said that this would arrive sometime in the 2030s or possibly the 2040s. Like, a lot of these people were imagining a world in which, within the next two decades, we will have a flock of literal satellites around the sun capturing its energy.

• 31:50Casey Newton?

  Yeah, and I do think it speaks to, you know, maybe a minor flaw in the conference, Kevin, which is despite the fact that they wanted to bring together all sides of the AI debate, as you said, they really mostly found a bunch of people who thought superintelligence is going to be here very soon, and also were very, very nervous about what that means.

• 32:10Kevin Roose?

  Yeah, there were several different kinds of clusters of people at this conference. I would say the biggest by far were the sort of AI safety people, the people who are sometimes described as doomers or rationalists, the people who have been warning about the risks of advanced AI for years. But I did see, for example, that one of the authors of the book AI Snake Oil was there and had an interesting sort of debate with former hard-fought guest Daniel Cocotello, formerly of OpenAI, about sort of whether, you know, his position was that basically AI is kind of just a normal technology, and that it won't have these sort of outsized effects on the world, at least in the timeframes that a lot of people

• 32:52Kevin Roose?

  at this conference are thinking. So that view was represented, but I would say it was the minority view.

• 32:57Casey Newton?

  Well, and it's, I also attended that debate and it sort of felt like the debate was like, superintelligence will get here in 2026 versus superintelligence will get here in 2028. Like, that sort of felt like the debate.

• 33:08Kevin Roose?

  Right, but that is, I think that is a useful calibration for people who are not spending a lot of time in the world of AI, is that even the pessimists, the skeptics, who are in part of this world, believe that it will only be a decade or two at most before the world looks radically different as a result of AI.

• 33:29Casey Newton?

  Yeah, I think that's right. Well, so, do you want to maybe talk about a thing or two that you saw at the conference that maybe taught you something new or maybe made you see this world a little bit differently?

• 33:40Kevin Roose?

  One thing that really struck me is that you really are starting to see evidence that people are changing the way they live their lives based on this, I think, quite sincere belief that the world is about to look radically different. I met one person at this conference who said that they had stopped saving for retirement because they believe that they, you know, in the post AGI world, money will have no meaning. And so what does it matter if you have a 401k 25 years from now if the robots are, you know, serving all of our material needs?

• 34:11Casey Newton?

  You know, that's, that's what I would say if I ever lost my life savings by gambling, Kevin. I would say, well, look, AGI is almost here. What does it matter?

• 34:19Kevin Roose?

  Yeah, it's true. I met another person who said that they are trying to get like really fit and really healthy and really hot because like, when intelligence is just like in a machine somewhere, like, being smart won't be a status indicator anymore. Like, we'll kind of go back to a very superficial way of like evaluating status, because the AIs are going to be so much smarter than us that all that'll matter is your, like, puny human body.

• 34:45Casey Newton?

  Wait, no, this is the worst news that you could possibly tell me, Kevin. I can't be out there competing on my looks. It's not going to work out.

• 34:53Kevin Roose?

  I think you're gonna be great.

• 34:55Casey Newton?

  Well, maybe AI will, I don't know, be able to give me some kind of facelift or something. I don't know.

• 34:59Kevin Roose?

  But like, what kind of conversations were you having? What stuck out to you about this event?

• 35:03Casey Newton?

  Yeah, well, so, I mean, one thing I found interesting was just how many of them are already thinking about geopolitical conflict, right? So, you know, we love to, on this show, talk about, well, yeah, what will happen if we have an AI that cures cancer or like makes the average worker much more productive, hopefully without costing them their job. These folks are thinking about like, well, if the United States is on the precipice of reaching superintelligence, how is China going to react to that? And could that possibly trigger some kind of live war? And there were some war gaming exercises at the conference where people were trying to map that out. And this is just a dimension of this stuff that, you know, again, speaks to how seriously everyone's taking it.

• 35:45Kevin Roose?

  Yeah, this was my favorite session at the event. Actually, I didn't participate in it, but I watched most of it. They had like a three hour long, what they called a tabletop exercise, where, have you ever been, did you do Model UN in school?

• 35:59Casey Newton?

  No, but I played Monopoly.

• 36:00Kevin Roose?

  Okay, so it's sort of like Model UN. Basically, you assign people roles. And so one person was the Chinese government and one person was the US government and one person was OpenAI and one person was Elon Musk and one person was, you know, the representative of the press and the public. And they did this exercise where they would basically walk through various sort of checkpoints in the year 2027 and figure out what AI was capable of doing at that point and then have everyone in their roles kind of react to that and figure out what they were going to do. And some of the interesting things that happened in that were things like international espionage. You know, at some point, the person who was representing the role of the Russian government

• 36:45Kevin Roose?

  tried to steal the weights of one of the leading AI models. The Chinese government was also conducting espionage to try to steal the sort of models that the US AI companies were building. And then on the sort of domestic front, the person who was playing the role of the US government at a certain Elon Musk in this exercise convinced Donald Trump in the year 2027 to nationalize all three of the leading American AI labs, so Anthropic, OpenAI, and Google DeepMind, and basically install himself as the head of it. And all of this is fictional, all of this is a simulation, but it's the kind of thing, these are not sort of like random people playing this tabletop exercise.

• 37:28Kevin Roose?

  These are people who are really in the room when some of these issues are being discussed. And the fact that that did not seem implausible to them as a potential thing that could happen two years and change from now just really made me sit up straight and pay attention and really made me hope that lawmakers in Washington who actually do have to deal with AI policy on a daily basis, not in a simulated way, are going through these same kinds of exercises.

• 37:54Casey Newton?

  Yes, absolutely. And for what it's worth, I find that scenario totally plausible, right? Assuming that Musk and Trump's relationship endures in the state that it's in for another two or three years. Nobody has a savior complex like Elon Musk. And if we're truly on the precipice of reaching superintelligence, it's hard to imagine the US government under any administration just telling private companies, yes, sure, you can wield superintelligence however you see fit. So again, this is why I was glad that we got to spend a couple of days thinking about, okay, what should the relationship be between government and hype super powerful AI?

• 38:29Kevin Roose?

  Yeah. Casey, what was the most memorable thing that you attended during this conference?

• 38:35Casey Newton?

  That would have been a session called If Anyone Builds It, Everyone Dies, which was hosted by Eliezer Yudkowsky. Eliezer is sort of the original doomer. For a couple of decades now, he has been warning about the prospects of superintelligent AI. His view is that there is almost no scenario in which we could build a superintelligence that wouldn't either enslave us or hurt us, kill all of us, right? So he's been telling people from the beginning, we should probably just not build this. And so you and I had a chance to sit in with him. People fired a bunch of questions at him. And, you know, we should say like, he's a really polarizing figure.

• 39:18Casey Newton?

  And I think is sort of on one extreme of this debate. But I think he was also really early to understanding a lot of harms that have bit by bit started to materialize. And so it was fascinating to spend an hour or so sitting in a room and hearing him make his case.

• 39:34Kevin Roose?

  Yeah, and I think to people who aren't sort of steeped in this AI world, like some of what we're describing may sound quite extreme. And even to some people who are in the AI community, like these are, these are some voices that are considered fairly extreme in these debates. But these are also people who have been quite influential, like Eliezer Yudkowsky, you know, was, was part of why, you know, the original investment in, in DeepMind was made back when that was an independent company. He, he sort of introduced Dennis Hassabis to Peter Thiel. And these are people who, despite the polarizing nature of their views, have been quite influential inside the world of people who have been thinking in this way for a long time.

• 40:16Casey Newton?

  Yeah, my case for taking these folks seriously, Kevin, is that this is a community that over a decade ago, started to make a lot of predictions that just basically came true, right? They started to look at advancements in machine learning and neural networks and started to connect the dots. And they said, hey, before too long, we're going to get into a world where these models are incredibly powerful. And all that stuff just turned out to be true. So that's why they have credibility with me, right? Everything they believe, you know, we could hit some sort of limit that they didn't see coming. Their, their model of the world could sort of fall apart. But as they have updated it bit by bit, and as these companies have, you know, made further advancements and they've built new products, I would say that this model of the world has basically held so far.

• 40:58Casey Newton?

  And so if nothing else, I think we have to keep this group of folks in mind as we think about, well, what is the next phase of AI going to look like for all of us?

• 41:05Kevin Roose?

  Yeah, I think the thing that I came away feeling is both, you know, respect for the people who've been thinking about this stuff for a long time and who have remained relatively consistent on it. But also, I feel like the AI safety folks who are sometimes called the doomers, I think they have a messaging problem. And I think that they have, they have a big task of persuasion ahead of them. Because if you have been warning about the risks of AI for years and years, as some of these folks have, I just think at a certain point, people start to tune you out, right? There's the kind of crying wolf problem in AI safety. You know, there were people, for example, who thought that GPT-2, the sort of precursor to ChatGPT, was going to be quite dangerous.

• 41:49Kevin Roose?

  And it turned out GPT-2 was not dangerous, right? People had the same sorts of concerns about GPT-3 and GPT-4. And then there was that, you know, there were these sort of, after ChatGPT came out, there were all these open letters saying we should pause AI development because this stuff is moving way too fast. And I, I just think these, these folks need better evidence for the kinds of claims they're making. Because people who are not steeped in this world, who are not going to these sorts of events, they look around, they see some harms from AI, but it's not existential harms, right? It's, it's people getting scammed, it's deep fakes showing up on their social media pages. It's like annoyance and, and a little bit of sort of worry about their jobs being taken away. It's not Dyson's fears by 2040, right?

• 42:31Casey Newton?

  Right. You know, Kevin, I actually thought you had a great idea for how this group of folks could fix this. And that was to just make a really good sci-fi movie about AI doom. So can you make the case for somebody making a good AI doom movie?

• 42:46Kevin Roose?

  Yeah, so this was sort of a half joking suggestion that I, I made, I think to Eliezer or some, some other sort of eminence in the world of AI safety. But my take on this has been informed by just going around, talking with people about AI for the last couple of years and realizing that so many of people's ideas about AI are derived from Hollywood, right? They are derived from the Terminator, from Skynet, from Ex Machina, from her.

• 43:12Kevin Roose?

  As a culture, the way we get our ideas about the future And whether we should be excited or scared about what's coming is often through film and through pop culture. And I just haven't seen a lot of really good attempts to sort of distill the scenarios that the sort of more pessimistic folks are worried about into a really compelling story that could actually attract a mainstream audience. Because I think if you took a mainstream audience and like brought them to this conference, people would be like, what the hell are any of these people talking about? What is a Dyson sphere? Why are we talking about paperclip maximizers? And like, there's just so much jargon and inside talk among these folks who have been talking to and with each other for many, many years, that it just sort of feels impenetrable and you don't even really know where to start.

• 43:57Kevin Roose?

  But if you had a movie that was, you know, made by people who actually understand the trajectory and pace of this technology, I think that could be quite convincing.

• 44:07Casey Newton?

  I think you're so right about this. I think that sci fi continuously shapes and reshapes the conversation in tech. You know, the movie Her came out a decade ago and it is still a touch point, it seems like for everyone at OpenAI in particular, as they work on ChatGPT. So I think a movie that really sort of sketched out a still somewhat fantastical, but increasingly plausible scenario for how a superintelligence might wind up being extremely harmful, I think would be super useful. And I think it would just be really kind of entertaining. I mean, as we were sitting in Eliezer's talk, he sort of sketched out one of these scenarios. And he was sort of like, you know, look, if you think you're just gonna like run away to your cabin, like it's gonna be able to find you in the cabin, you know?

• 44:53Casey Newton?

  And, and you know, maybe it'll just sort of take you out with a drone strike or maybe you'll just fall sick one day and you'll start coughing and you'll be dead a few days later. I mean, and as he was like describing that, I'm seeing the movie play out in my mind, right? Of like what that would feel like. But I will say, Kevin, that there has been an AI doom movie that was made this year. And I wonder if you've heard of it?

• 45:13Kevin Roose?

  Which one?

• 45:13Casey Newton?

  Have you heard of Afraid?

• 45:15Kevin Roose?

  I have not seen it, but I haven't heard about it in part because I believe that a quote from my conversation with Bing Sydney is like the opening scene of the movie. And so I did get some texts from people saying, have you seen this thing? I have not.

• 45:28Casey Newton?

  Wow. Your work has inspired so many horrors. It's really, it's really remarkable.

• 45:33Kevin Roose?

  Some people have said that the work itself is a horror film.

• 45:36Casey Newton?

  I think your editor said that to me once.

• 45:39Kevin Roose?

  I get that a lot.

• 45:41Casey Newton?

  So Afraid, and it's stylized so that the AI in Afraid is capitalized. It's basically about an experimental AI that takes over a house. And then, of course, it, you know, goes haywire and starts wreaking havoc. So, you know, I haven't seen it yet, but it is on one of the big streaming services. And so I'm going to check that out this weekend and see if maybe this is the AI Doom movie you've been requesting.

• 46:07Kevin Roose?

  I mean, what the one sort of backdrop for all this is that like the the general public actually is quite pessimistic about AI. There's been some data recently that from Pew that shows that the the number of Americans who believe that who are more concerned about AI than they're excited about it has risen sharply in the past couple of years. So people are quite worried. But I think if the if the optimists, the people who believe that AI is going to transform the world for the better, want that narrative to take hold among people who are sort of mainstream, you know, media consumers, they should start making some some movies where AI is the hero. And that I think could convince people of that, too. It's it doesn't make for as good a film, probably, because people like, you know, doom and horror and thrill.

• 46:48Kevin Roose?

  But I think the role of fiction is underappreciated in shaping people's opinions about technology.

• 46:54Casey Newton?

  Yeah, and and I would add that while, you know, a sort of utopian movie about AI might feel less compelling just because, you know, what is the the drama or the conflict there? It's important to note that the people who are at this conference, even though they're super concerned about it, they're working on AI because they think it can do massive amounts of good, right? They do believe that it can cure cancer and all other human disease, right? It's sort of, you know, you name a human challenge, and they think that superintelligence is going to be the way that it gets fixed. And so if there is a way of, you know, making the like how we cured cancer, you know, using AI and everything that happened afterwards kind of movie, well, that could be kind of good.

• 47:31Kevin Roose?

  So after this Curve Conference, are you more optimistic or more pessimistic about AI? Do you feel like we are headed toward the the AGI era that people have been warning about for years? Or do you think these fears are somewhat overblown?

• 47:46Casey Newton?

  I mean, the way that I would put it, Kevin, is I came out of that conference just taking AI more seriously, which may sound like a funny thing to say, given that it is really a primary topic of this podcast. We're interviewing people every other week about it. And yet, sometimes I still feel like I haven't been totally committed to the idea that superintelligence could arrive within the next five years or so, and starting to grapple with what that might mean. So I left there sort of committed to learning more about this technology, sharing more about what I'm learning, make it sort of more of a primary focus of my reporting, and absolutely taking the risks seriously. But taking some of the positive outcomes and possibilities seriously, too.

• 48:31Casey Newton?

  How about you?

• 48:32Kevin Roose?

  I don't think it really changed my view of AI and AI's trajectory, but I think it's useful to have these kind of gatherings, if only to sort of bring people face to face. I think often, these conversations take place on social media, where they can become very bitter and very polarized. I don't like how polarized AI discourse has become. I think these are real challenges, and we need to assume that people are trying to solve them in good faith. And what I came away from the Curve thinking and just being reminded of is just that many of the people who work at these companies and who work on this technology are quite sincere. Like, they might be sincerely wrong. Like, they might be wrong, right? We should just say that.

• 49:15Kevin Roose?

  Like, they might, these fears might be fantastical. They might never materialize. But I don't think I believe anyone who doubts the sincerity with which these people hold these views. It's not just stuff they say when they're raising money. It's stuff they're talking about in the comfort of their homes, in these more private venues, around the fire at night at the Curve Conference. Like, they really do take this stuff quite seriously. And many of them are starting to act on it in ways that I think would surprise people who just think they're saying this to get a rise out of people.

• 49:47Casey Newton?

  Yeah, well, I think we're gonna have a big 2025, Kevin.

• 49:51Kevin Roose?

  Yes, and that may be the last year.

• 49:55Casey Newton?

  Yep. So if there's any loved ones you'd like to say some goodbyes to, really make the most out of this Christmas.

• 50:04Kevin Roose?

  Yeah.

• 50:08Kevin Roose?

  When we come back, it's gift guide season. We'll talk about why the internet is plastered with gift guides and why you should get your family this year.

• 50:29Speaker 00

  Support for this podcast comes from Dell Technologies. It's time to put AI to work. Is your business ready? The challenge isn't getting started with AI. It's scaling beyond pilots while maintaining control of your data, governance, infrastructure, and costs. Dell AI Factory provides the integrated foundation you need to scale AI securely and efficiently with one trusted partner. Dell, AI on your terms. From idea to impact. Learn more at dell.com/yourwaytoAI. That's dell.com/yourwaytoAI.

• 51:03Speaker 04

  AI spend is growing fast, but most businesses still can't see exactly where that money is going. Ramp gives businesses visibility and control across AI providers with real-time guardrails by team and model, helping ensure spend goes to the highest ROI areas. Stop guessing where your AI spend is going. Get started with Ramp. And if you're new to Ramp, you can get $250 off at ramp.com/hardfork. Go to ramp.com/hardfork. Ramp.com/hardfork.

• 51:35Speaker 06

  Today's AI builders need infrastructure that scales with their ambitions. iRan is a vertically integrated AI cloud platform, delivering data centers, compute, and software for AI training and inference, underpinned by its expansive portfolio of land and grid-connected power in renewable-rich regions globally. iRan's AI Cloud gives direct access to Nvidia GPUs, scalable from hundreds to thousands. For builders, that means direct access to AI compute, rapid scaling, and predictable pricing. iRan, AI Cloud for the next big thing.

• 52:10Kevin Roose?

  Well, Casey, it's the most wonderful time of the year.

• 52:13Casey Newton?

  Oh, you mean OpenAI's recently announced 12 days of shitmas?

• 52:16Kevin Roose?

  No, I'm talking about the holiday season. And on the internet, increasingly, in the year of our Lord 2024, that means the season of gift guides. Now, Casey, are you a big gift guide consumer?

• 52:31Casey Newton?

  Well, you know, I didn't used to be, Kevin. And then something happened. And now it's basically not possible to visit a website in December that does not immediately show you its gift guide and urge you to make purchases based on it.

• 52:43Kevin Roose?

  Yes, I have noticed this too. Like everywhere you look, there are gift guides. They are inescapable, starting like a week or two before Thanksgiving and continuing through Christmas. It is sort of the version of like the Christmas trees going up in the malls. Like, that's how you know it's Christmas season, is when Starbucks starts selling the eggnog latte and the gift guides start appearing on every website on earth.

• 53:04Casey Newton?

  Yeah, gift guides are the pumpkin spice latte of December.

• 53:08Kevin Roose?

  Totally. So as we were thinking about what to do on the show this week, knowing that the holiday season is fast approaching, we thought we would do sort of a deep dive into the gift guide economy, because it is a surprisingly interesting story why gift guides have sort of taken over, at least the parts of the internet that we visit. And it is actually a major part of the media story over the past few years, is this shift away from ad revenue and toward affiliate revenue. So we'll get into all of that. And then we actually have some of our own a hard fork gift guide for our listeners. But Casey, just talk to us about what the sort of economics of the gift guide market are. Why are these things so freaking popular?

• 53:48Casey Newton?

  Yeah, so this is what makes them interesting to me, is that gift guides are a story about how the internet evolved. So you go back 10 or 20 years, and publishers were able to make a pretty good business just out of display advertising, maybe a little bit of what they call programmatic advertising, which is when advertisers buy ads through auctions.

• 54:07Kevin Roose?

  This is banner ads you're talking about.

• 54:08Casey Newton?

  Yes, banner ads, for example. And then Facebook and Google start running away with the advertising game. And yeah, there are some other kind of big competitors. Amazon winds up building a pretty big digital advertising business, too. But for the most part, the publishers are just having a harder and harder time competing. As they start looking around for what are the other sources of revenue out there? And they land on affiliate marketing. Because many makers of consumer products are happy to split a little bit of the profits with an affiliate, if that affiliate can make a sale. And so websites just lean into this in huge numbers. And eventually, that brings you to the gift guide, which is just a list of things you can purchase.

• 54:53Casey Newton?

  And for the most part, all of those are going to have an affiliate link in them. And the more things that websites can get you to buy, the more money they can make. And from their perspective, it is basically free money, because nothing is easier to write than a gift guide. And if it prints cash for you, well, then great. Maybe you can keep the lights on for another few months.

• 55:12Kevin Roose?

  So Casey, just explain for people who may not be paying close attention to the economics of digital publishing, like, how does affiliate sales work?

• 55:20Casey Newton?

  Sure. So when you click on a link that is in a gift guide, that link is going to have some extra characters at the end that tie it to the publication that you are clicking from, right? So it's going to probably leave a little bit of a cookie on your computer, and it just sort of tells your computer, hey, you know, this publication is linked to the sale. Then, once you complete the sale, the maker of that good is going to kick some of that revenue back to the affiliate. And this is now increasingly what makes the world of the internet go around.

• 55:52Kevin Roose?

  And what is the size of the kickback? Like, are we talking a few cents? If I buy something that costs a lot of money, like a new TV from an affiliate link, are they getting, you know, $100 from that? Like, what kinds of kickbacks are the publishers getting?

• 56:07Casey Newton?

  So I read an interesting blog post on this subject recently. The blog is called Shop Rat. It's written by Amelia Petrarka. And she wrote that some brands will offer affiliate partners 10% or even more than 20% to link to their site. So some of those affiliate percentages can get quite high. And she notes that those are the brands that you are likeliest to see in a lot of these gift guides, right? Because there is an economic motive that sort of intersects with the editorial prerogatives. And, you know, look, if you can make 21% on one pair of jeans, you might be likelier to recommend that over the pair of jeans that only kicks you 2%.

• 56:48Kevin Roose?

  That's really interesting because it might actually be that the sort of affiliate structure privileges certain merchants over others. Do we know how widely that, like, how much of this stuff is corrupt, basically, by these affiliate fees?

• 57:02Casey Newton?

  Well, here's the thing. I'm sure that every single maker of a gift guide would tell you that these things that they are recommending come from the bottom of their heart, and they don't even look to see what the percentage is that they're getting. And I think some brands are more scrupulous than others. But one of the reasons why we want to talk about this today is some of this stuff is really unscrupulous, right? You know, journalism is kind of a weird profession where we try to keep the people that write the articles mostly pretty far away from the people who are selling the ads. But one of the things that that's supposed to do is to buy us some credibility, right? Because if you're writing about the new iPhone, particularly if you work at the New York Times, you're not getting a kickback, you know, from the, from Apple for every time, you know, somebody buys an iPhone.

• 57:44Casey Newton?

  But as we've moved into this new phase of the internet where there's less advertising to go around, more and more publishers are saying, you know what? Why don't we have a gift guide, not just for Christmas, Kevin, but for basically every single day of the year?

• 57:58Kevin Roose?

  And of course, we should acknowledge that The New York Times owns Wirecutter, which is our product recommendation part of the business that uses affiliate links and does gift guides and all that, too. But they have a whole fascinating section on their site called "How Wirecutter Makes Money," where they outline how they prevent this kind of affiliate marketing world from bleeding into their product recommendations. And basically, what they say is, after we give our writers and editors a set of products to test, after they come back, they write their review, they tell you which kitchen gadgets to buy. At that point, and only at that point, do they kind of send it to the people who decide which links to put on those products, where they might consider, does the retailer have

• 58:39Kevin Roose?

  enough inventory to satisfy all the orders for this stuff? What are the shipping costs like? And they also do take into consideration affiliate rates. But they are basically saying, the affiliate rates are not determining which products we review and how they review them. They're just determining, do we send you to Amazon, or Walmart, or Target, or some other website?

• 59:00Casey Newton?

  All right. Well, that's fair. But in the spirit of fairness, Kevin, I would also say that the Wirecutter reviews a lot more products than it used to. So I'm not saying there's anything nefarious about that, but I do think it speaks to the way that the internet has been transformed by the need for publishers like The Times and others to just find big news sources of revenue. And increasingly, that means recommending lots and lots of products.

• 59:21Kevin Roose?

  Yeah. Now, something that I've been doing this holiday season, because I am AI-pilled, is that I've been asking chatbots for product recommendations for specific people in my life. Like, I'll say, I'm looking for a present for my brother. Here's some facts about him. What do you think? And I know that it's going out, and it's pulling from all these gift guides behind the scenes. But is there any fear that the advent of AI chatbots and AI chatbot shopping could disrupt the kind of system of affiliate marketing and sales and gift guides that we see on the internet today?

• 59:55Casey Newton?

  I think so. I mean, first of all, you know, when I ask AIs to recommend a product, it will always suggest books about how to free super intelligences from their computers and sort of release them into the world. So that's just something that I find sort of unsettling. But no, you're right. This does really threaten the model because, you know, up until, let's say, the present day, a lot of us might search for something like the best laptop and we'll go to the Wirecutter, we'll go to The Verge, and they've done a really, you know, great, great investigation of that. And they'll show their work and their recommendation has a lot there to support it. But what an AI can do is it can just look at every single website's best laptop and sort of average those out and say, well, most people say that you should buy this particular MacBook, let's say. And then the

• 1:00:38Casey Newton?

  search engine can insert its own affiliate link. And even if that company hasn't done any of the work, aside from statistically averaging every other website, they can hypothetically then reap all of the rewards of that and cut The Times, The Verge, and everyone else out.

• 1:00:51Kevin Roose?

  I mean, Perplexity, the AI search engine, did just release a shopping tool that allows you to find, for example, I'm looking at it right now, it has the best headphone recommendations. And then it gives you a link to some products. And if you click on those products, the link actually is to Perplexity, and it does have their, what's called a UTM code in the URL. So it does appear that in the case of Perplexity, they are sort of disrupting this affiliate pipeline.

• 1:01:22Casey Newton?

  Yeah. And having interviewed this year of Perplexity, let's just say, that's what I would expect.

• 1:01:26Kevin Roose?

  Yep. So, okay, that is the economics of gift guides, but we promised our listeners that we would come up with our own gift guide this year, the Hard Fork Gift Guide. And so we both brought some recommendations of things that we think our listeners would like to maybe get for their loved ones this year.

• 1:01:42Casey Newton?

  First of all, we should say, are we earning any affiliate fees on this?

• 1:01:46Kevin Roose?

  I hope so, but I don't think we are.

• 1:01:48Casey Newton?

  No, we're not. These are, these are Hard Fork freebies.

• 1:01:50Kevin Roose?

  Yes. We're not what's called good at business. So we're not, we're not making any money on this. This is purely for the love of the game.

• 1:01:56Casey Newton?

  We have a bit corrupted by the internet. In this, in this specific way.

• 1:02:01Kevin Roose?

  But I will say that three of these businesses are owned by a shell corporation that I can't talk about. So my first gift guide recommendation this year is for the home cook. I am a home cook and the best thing that I have added to my kitchen arsenal this year is a Zoji Ruchi rice cooker. Casey, do you have a rice cooker?

• 1:02:20Casey Newton?

  Not only do I have a rice cooker, Kevin, I have a Zoji Ruchi rice cooker.

• 1:02:23Kevin Roose?

  I mean, this thing has brought me so much joy. I got it used on Facebook Marketplace for $60 and I now have rice many times a week. This thing is so easy. You put the rice in, you put the water in, you don't have to really have to measure, you just kind of eyeball it, push the button, it plays a cute little song and then out comes your perfect rice later. It is truly a magnificent gadget.

• 1:02:47Casey Newton?

  Now Kevin, some people might say, why buy something for just a single purpose, right? I already have a pot. What do I need a rice cooker for? Is that, is that really worth the counter space?

• 1:02:55Kevin Roose?

  Yeah, so this is the most common objection that people, including my wife, bring up when I talk about my rice cooker. And I will say, it has earned its place on, on my counter because I now eat rice so much more than I used to. I promise you, it is worth the investment if you are a person who enjoys rice. The first thing I did when I went, when I got my rice cooker, I went to the grocery store and I got 15 pound bags of several different kinds of rice. So now I'm prepared for the apocalypse. I'm never going to run out of rice.

• 1:03:24Casey Newton?

  By the way, this segment was brought to us by the National Rice Council and we want to thank them for their support. Well, Kevin, here is an item I would like to, to recommend for what you might call the extravagant book lover. Because we all know at this point that you can gift a Kindle book, okay? And we all know, or maybe less of us know, that you can gift an audiobook through audible.com. What you might not know is that if somebody has a Kindle book and an audiobook for the same book, a feature is enabled called WhisperSync. Are you familiar with WhisperSync?

• 1:03:59Kevin Roose?

  No.

• 1:04:00Casey Newton?

  Here's what it means. It means that you can be walking down the street on your commute, listening to a book, and then when you get back home and you open up your Kindle. Your iPhone, the place that you were in the audio is synchronized to the text. And you can go back and forth. And I have been able to enjoy so many more books this way because now reading is a multimedia experience. And if you really want to have an experience that will take you back to when you were a child and maybe a parent was reading to you, you can actually open up your Kindle and listen to the audio at the same time and it'll highlight the words for you. And it's almost like your mom is reading you a bedtime story. And this has truly become my favorite way to read a book.

• 1:04:43Kevin Roose?

  Huh. And is WhisperSync like an add-on to a Kindle? Is it a separate subscription? Or how does it work?

• 1:04:49Casey Newton?

  This is just technology that Amazon is built. So I know this would be a great time to recommend a local independent bookstore. And of course, I recommend that as well. But if you want a sort of high-tech book experience, WhisperSync is just baked into the Kindle software. And I'm telling you, it's just a really fun way to read.

• 1:05:08Kevin Roose?

  This has been a very effective gift guide so far for convincing me that you know how to read or do read books with irregularity, which I don't believe.

• 1:05:15Casey Newton?

  All right. Give us another recommendation, Kevin.

• 1:05:18Kevin Roose?

  Okay. This next one is for the gear dads out there. The best piece of technology that I added to my parenting life this year was a cargo e-bike. So I, you know, I've seen these things around and by cargo e-bike, I just mean an e-bike that is sort of elongated and has sort of a back rack on which you can put some groceries, maybe. A lot of people put child seats on them so that their kids can ride on the back of the bike. And these things have been around for several years. I've seen people with them. I've been a little sort of interested and curious. But this year, I finally took the leap and got myself a cargo e-bike, which, you know, the advantage of the e-bike part is that if you're going up a hill or you're carrying

• 1:06:01Kevin Roose?

  a heavy load, like these things are quite heavy and so they can be quite hard to pedal. But with the e-bike part, I just zip around town. It's sort of become our family's second car. My kid loves riding it. And now there's this sort of like cult of e-bike people that I feel like I've joined. And so I highly recommend a cargo e-bike. The specific one I have is called the Aventin Abound, which I really enjoy. But there are lots of other models out there and the prices on these things are coming down. They're still a little bit pricey, but compared to a second car, pretty affordable.

• 1:06:33Casey Newton?

  Okay, my next recommendation is for anyone who travels, which I think is for most of you. But I'm telling you, you want to make your traveling life easier. Here's what you do. You gotta get one of these tech organizer cable bags. Kevin, have you seen these?

• 1:06:47Kevin Roose?

  No.

• 1:06:48Casey Newton?

  Okay, so this is a little zippard pouch that you put into your carry on bag. And into this pouch, you put your lightning cable, your USB cable, the sort of wall charger, you know, sometimes called wall warts. And what I want, here's what I want you to do. For so long, every time I had to travel, I'm ripping all of my cords out of the wall and I'm wrapping them up and I'm shoving them in my bag. And I'm praying that I remember to bring all the right ones. And I'm praying that I didn't leave them at the hotel on the way back. And I'm just here to say, if you're listening to Hard Fork, it's time to upgrade your life and respect yourself and actually just get a duplicate cable for everything that you need. And you're going to put it in this bag and you're going to put that in your carry on.

• 1:07:32Casey Newton?

  And then you are going to be so much happier in your life.

• 1:07:35Kevin Roose?

  This sounds like a great idea. I personally probably need one of these. I'm a person who travels with a lot of cords and I spend like minutes at my hotel wherever I'm going, like just sort of untangling the cords from each other.

• 1:07:45Casey Newton?

  Exactly.

• 1:07:46Kevin Roose?

  And this sounds like it would solve my problem.

• 1:07:47Casey Newton?

  Exactly. And this is the thing. Here's what I like about this. You could just truly do it for every single person in your family. You get them a little organizer. If you're an Apple family, you get them a couple of extra lightning cables. You know, you're an Android family, you get them a USC cable. And this is just something that they are truly to be able to. A what?

• 1:08:06Kevin Roose?

  USB C?

• 1:08:07Casey Newton?

  What did I say?

• 1:08:08Kevin Roose?

  You said USC.

• 1:08:09Casey Newton?

  No, I mean, get them a USC cable to support the University of Southern California.

• 1:08:13Kevin Roose?

  Okay, good.

• 1:08:15Casey Newton?

  Or whatever college you went to, see if they have a cable. So this is just something that you can do. And this is kind of the evolution of something that I used to do for my family, which was just get everyone a battery pack for their phone. Because you know, phone batteries used to die. You know what everyone loves? Having an extra phone battery. You can just charge it during the day. Well, now I'm telling you, the upgrade is just get everyone a duplicate set of their favorite cables and put it in a little bag. They're going to be so much happier.

• 1:08:39Kevin Roose?

  Love that. Okay. My next gift guide suggestion is for the nostalgic nerd. This is something that is going on my list this year. I do not actually own any of these, but I am a fan of them. And if any of my family members are out there listening, this would be, let's just say if it showed up under the tree, I wouldn't send it back. These are from a company called Ex Reart, and they are making these very lovely framed old pieces of technology. So you can get, for example, a framed first edition iPhone in which they have basically disassembled the iPhone and laid out all the parts on a beautiful mat

• 1:09:21Kevin Roose?

  with diagrams showing like what all of the components of the first iPhone are. And they've got these for things like the original Nintendo Game Boy, for the Apple Watch. They've got these for even older phones, like the old sort of candy bar Nokia phones that everyone used to have. They look great. They come in a sort of box frame and you can put them on their wall. So that is something that I am, if no one gets it for me for Christmas, I'm going to buy my own this year because I really like the way they look.

• 1:09:51Casey Newton?

  I mean, if no one gets it for you, Kevin, I will get it for you.

• 1:09:54Kevin Roose?

  Oh, thanks.

• 1:09:55Casey Newton?

  I would be so sad if after broadcasting this to the entire world, no one got this for you, I would step in.

• 1:10:04Kevin Roose?

  Santa, if you're listening.

• 1:10:06Casey Newton?

  Now, I think, Kevin, so far, these picks have hit a lot of people's budget. Not everyone's budget, but you know, I think a lot of our listeners would be able to afford these. But I think we should also, you know, be real about the fact that a lot of our listeners are profoundly wealthy.

• 1:10:21Kevin Roose?

  Yes, we have many billionaires in the Hard Fork audience. And what should they get for their family members?

• 1:10:28Casey Newton?

  Well, I'll tell you about a circumstance that I found myself in, Kevin, which was that when I bought my house, it had a keggerator in the kitchen.

• 1:10:35Kevin Roose?

  No.

• 1:10:36Casey Newton?

  Yes, there was a sort of door that you would open and there were hookups for, you know, kegs of beer. And while certainly I enjoy a beer, I thought it would be foolish for me at this age to just have a keg of beer at my house. Since I do not plan on throwing any frat parties, you know, in the near future. So I talked to a friend and she said, you know, you could actually use the same setup and you could put a little bit of wine on tap and you could put a little bit of kombucha on tap. And so for like a hundred bucks, I bought like keg of kombucha and a keg of wine. And now if you come over to my house, Kevin, you can enjoy some blood orange kombucha or a glass of rosé on tap. And let me just say, when someone comes into your house and you say, can I get you some wine

• 1:11:21Casey Newton?

  or kombucha on tap? The double take they give you is worth everything that you spent to have that on tap.

• 1:11:25Kevin Roose?

  This is the most decadent thing I've ever heard about you. You are now Cleopatra in my eyes. You, you, you are, you are the wealthiest man in San Francisco.

• 1:11:38Casey Newton?

  I, I, I truly am not, you know, in San Francisco, I think I'm, you know, solidly have San Francisco middle class. But the important thing here, Kevin, is that this is a really fun and cool thing. And I think there are honestly pretty inexpensive ways to do this. A lot of people have wine on tap in their house and it's just really fun.

• 1:11:54Kevin Roose?

  Okay, so if you're rich and you want to develop a drinking problem, that is the gift for you.

• 1:12:02Kevin Roose?

  Casey, I have a, I have a lower priced gift guide recommendation too. And this one is for the doom scroller. If you spend too much time strapped to a screen, watching the headlines roll by and just making yourself miserable, I recommend that you throw more dance parties for yourself. This is something that I have started to do this year in my house. My, my kid is obsessed with this. And it's a nice, become a nice family ritual. And there are two things that I would suggest adding to spruce up your dance parties. The first is a 100 pack of glow sticks. These are these little sticks that you can snap, they glow. You can buy them on Amazon. They cost like 10 bucks for a hundred of them.

• 1:12:47Kevin Roose?

  They really make you feel like you are at Burning Man. And then there's also a company that makes something called 30 second dance party. This is a Portland based toy manufacturer. And this is literally just a giant yellow button that when you press it, it starts a 30-second dance party with some nice techno music. And so I will be getting these for all of my family members.

• 1:13:13Casey Newton?

  Is this, is it the same 30 seconds of music every time?

• 1:13:16Kevin Roose?

  You know, I don't know. I guess I'll find out when I get one. But I love a dance party, and there have been some studies, or at least one study that I saw, that showed that dancing is better for your mental health than basically anything else you can do.

• 1:13:29Casey Newton?

  Well, I believe that. I just, I can't get over the fact that you called me decadent before announcing that you were gonna buy 100 glow sticks at the same time. Even people who go to Burning Man don't buy that many glow sticks.

• 1:13:44Kevin Roose?

  Good luck with your shopping. And if you feel like supporting us on the Hard Fork Podcast, you can buy or gift a subscription to New York Times Audio. Or if you already have one, you can buy a subscription to Platformer, although I feel a little weird chilling for you.

• 1:14:00Casey Newton?

  But you know what? It sounded great. And it's something I wish you'd do a little more of on this show. I'm always telling people to read the New York Times.

• 1:14:28Speaker 00

  Support for this podcast comes from Dell Technologies. It's time to put AI to work. Is your business ready? The challenge isn't getting started with AI. It's scaling beyond pilots while maintaining control of your data, governance, infrastructure, and costs. Dell AI Factory provides the integrated foundation you need to scale AI securely and efficiently with one trusted partner. Dell, AI on your terms. From idea to impact. Learn more at dell.com/yourwaytoAI. That's dell.com/yourwaytoAI.

• 1:15:02Speaker 04

  AI spend is growing fast, but most businesses still can't see exactly where that money is going. Ramp gives businesses visibility and control across AI providers with real-time guardrails by team and model, helping ensure spend goes to the highest ROI areas. Stop guessing where your AI spend is going. Get started with Ramp. And if you're new to Ramp, you can get $250 off at ramp.com/hardfork. Go to ramp.com/hardfork. Ramp.com/hardfork.

• 1:15:34Speaker 06

  Today's AI builders need infrastructure that scales with their ambitions. Iren is a vertically integrated AI cloud platform, delivering data centers, compute, and software for AI training and inference, underpinned by its expansive portfolio of land and grid-connected power in renewable-rich regions globally. Iren's AI cloud gives direct access to Nvidia GPUs, scalable from hundreds to thousands. For builders, that means direct access to AI compute, rapid scaling, and predictable pricing. Iren, AI cloud for the next big thing.

• 1:16:08Casey Newton?

  Hard Fork is produced by Rachel Cohn and Whitney Jones. We're edited by Jen Poyant. We're fact-checked by Caitlin Love. Today's show was engineered by Daniel Ramirez. Original music by Alicia Bietup, Marian Lozano, Diane Wong, and Dan Powell. Our audience editor is Nell Galopoli. Video production by Ryan Manning and Chris Schott. You can watch this whole episode on YouTube at youtube.com/hardfork. Special thanks to Paula Schuman, Puywing Tam, Dahlia Dodd, and Jeffrey Miranda. You can email us at [email protected]. But note that we're not accepting gifts.

• 1:16:44Kevin Roose?

  Except wine kegs. You can send those to Casey.

• 1:17:03Speaker 04

  AI isn't just opening new business opportunities, it's changing the way businesses get work done. The questions they ask, the decisions they make, and the routine tasks they can hand off. And nobody understands that better than NetSuite. That's why they put together a free expert guide you need to grab today. Aligning for the Agentic era, how AI is changing everyday work. Download it and you'll see exactly how AI is reshaping work across every part of your organization and how to make sure you come out ahead. Get your free guide at netsuite.com/NYT. netsuite.com/NYT.