The Decoder team turns the tables on Nilay and makes him answer your burning listener questions in our end-of-year wrap up special. We also reflect on the year’s biggest Decoder themes, discuss some of the most popular feedback we’ve received, and tease what we have planned for next year. Links: Here we go: The Verge now has a subscription | The Verge How The Verge Works | The Vergecast Intuit asked us to delete part of this Decoder episode | Decoder What’s really behind Big Tech’s return-to-office mandates? | Decoder Rabbit CEO Jesse Lyu isn’t thinking too far ahead | Decoder Transparent Vice | The Verge UiPath CEO Daniel Dines thinks automation can fight the great resignation | Decoder Palmer Luckey, American Vulcan | Tablet A revolution in how robots learn | The New Yorker Credits: Decoder is a production of The Verge and part of the Vox Media Podcast Network. Our producers are Kate Cox and Nick Statt. Our editor is Callie Wright. Our supervising producer is Liam James. The Decoder music is by Breakmaster Cylinder. Learn more about your ad choices. Visit podcastchoices.com/adchoices
Support for Decoder comes from Stripe. Stripe is a payments and billing platform supporting millions of businesses around the world. Whether you run a startup or a well-established business, Stripe can help you reach more customers, hit your revenue growth targets and build a more profitable business.
The platform offers a suite of specialized features and tools to fast track growth, like Stripe Billing, which makes it easy to handle subscription-based charges, invoicing, and recurring revenue management needs. You can learn how Stripe helps companies of all sizes make progress at Stripe.com. That's Stripe.com to learn more. Stripe. Make progress.
Hello, and welcome to a very special episode of Decoder. I'm Kate Cox, senior producer at The Verge, and while Decoder is Nila's show about big ideas, making it all work is my problem. I'm here today with my co-producer, Nick.
Hello, this is Nick Statt, fellow senior producer on Decoder.
And we are here with Nilay Patel, editor-in-chief of The Verge, who is host of Decoder and also our boss. Hi, Nilay. I love being the guest.
Being the guest is the best. If any of you listening have ever thought about hosting a podcast, you don't want to do that. You want to be a permanent guest. This is my dream.
We're here today to talk about some reader feedback because it's our end of year show. It's our last show for 2024. And we are very excited about all the things we've heard from you this year. We have had a really busy year. We interviewed a lot of people. We published more shows this year than in any past year, thanks to our second episode that we launched back in February.
It's been a lot of work and a lot of fun. So we wanted to take a second to look back on some of our favorite themes, address your most common feedback and talk a bit about what's next.
Yeah, and I thought the best way that we would do this is we would just grill Nilay with a bunch of questions. So yeah, in this episode, Nilay is the decoder guest. We ask listeners to send in some of their burning questions to answer. We also have a huge collection of emails from the past year. Yes, we really do read every single one.
We save a lot of them, and we pulled some of our favorites for this episode. But I think we're going to start with probably the most important question of all. It's the one, it's actually two related questions. very often. Some of you even have asked this question twice in the same week. We'll get an email one week, and then we'll get an email the next week with the same question.
And this is the big question. Nilay, why is Decoder not on YouTube Music?
Oh, man, we're coming out the gate hard. You know, I usually try to warm up the guest. I just want to be very clear about this. You don't dive into the controversy until the 20-minute mark.
We're going inverted pyramid.
Yeah. It's both a very simple answer and a somewhat complicated answer. The simple answer is money. The complicated answer is the way that we distribute the show is with a platform called Megaphone, which is owned by Spotify. That platform allows us to programmatically insert ads on download.
So if you download an episode three weeks after we produce it, it gets a new ad in it and we make the money from that ad placement at that time. That's very complicated. It's weedsy in the world of podcasting. But the thing to know is that YouTube Music doesn't let us play that game. It doesn't let Megaphone insert ads dynamically. You have to be in YouTube's system.
And you have to basically ship the audio to YouTube and let them do whatever they want with it. And the rates aren't the same. This is now, I think, very decoder, right? We're talking about platforms distributing content at different rates. We generally just want more control over our distribution. We want, obviously, more control over our rates.
We will find a way to play in the YouTube ecosystem. But all the way at the base of it is YouTube and Spotify aren't going to play nice with their different ad technologies. And we're kind of caught in the middle. And until we can find a way to not make less money on one platform than we do on another, we're just going to hold ourselves back.
That kind of brings us to a related question, which is why aren't we publishing full video versions of Decoder on YouTube? If that were, you know, people could get Decoder that way by just watching like they do on the Verge cast.
So there's only one Nick and there's only one Kate and there's only one Callie, our engineer, and we have to make two episodes of the show in audio every week. We try to do clips. You can see them on TikTok. They are fun on TikTok. They are really good marketing for the show on TikTok.
An interesting thing is we started the show in the pandemic and watching a bunch of CEOs just get immediately more comfortable with being on video has been fascinating. At the beginning of Decoder, they did not want to turn on their cameras. And now they assume... that they will get a two and a half hour video podcast to do whatever they want with on YouTube.
And that assumption I think is fascinating. It's a cultural change. We will probably find a way to use our archive and put the full video on YouTube in the future. If you just think back to the rates thing I'm saying, we've already made the money on those shows. So if we put them on YouTube, we're just getting more margin off the asset we've already made. This is now very decoder again.
But it's more work, right? We have to find a way to make that cost effective. So that's kind of the big project for next year is figuring out how to manage that investment and go to more places in a way that is sustainable. That's the big project for The Verge, quite honestly. You can see that we're changing our business model in various ways. But we know everybody wants us to do it.
We have the videos. We should figure out how to do it in a way that makes sense.
That ties into another technical and money and systems question that our readers ask a lot, which is why don't we have chapter breaks?
Megaphone, you can just blame Spotify. If we ever get Daniel Ack on the show, we will just ask him that question over and over and over again. I don't know if he knows the answer of how Megaphone works, but fundamentally, because we have dynamic ad insertion in Megaphone, we don't know how long the ads will be. And so it's very hard to insert the chapter blanks at the wrong time.
We've answered this question on the Verge cast side 5,000 times, and the solution is maybe the ads will just be weird, or there'll be like 10 seconds of silence. And a lot of people have told us that is acceptable, And then we do it, and then we get a lot of emails that are like, what's with this weird 10 seconds of silence? So we just have to find a way to make that work.
But everybody wants us to do chapters, and we'll figure it out.
We have a process question. It's another process question. This is from Justin Perez, who wrote us an email. He wrote, whenever I listen to some of the interviews, not just on this podcast, I tend to have one or more questions I wish I or the host could have asked the person being interviewed.
His suggestion for the podcast for Decoder is he would really like it if the audience got a chance to chime in on a question for the upcoming guest being interviewed. It's very logistically difficult, but Nilay, why can't we easily solicit reader questions and are there ways we really could do this in the future if we wanted to?
I would love to do this. I would love to be able to like put on blue sky or threads in real time. Like I'm interviewing the CEO. Like I'm just going to here's a real time feed of questions that I'm going to respond to. I think it would be really interesting. The problem is scheduling CEOs is very difficult. Most of the job is getting people to show up.
And like, I don't mean like show up like they don't want to do it or they're afraid of being grilled. I mean, literally, we need you to sit down in this chair at this time and turn on your microphone is a challenge because everyone's very busy. And so there's just a lot of times when I would say, hey, we're going to do this interview and then it moves a week or two weeks or falls off the schedule.
And the thing I hate doing the most is making promises I don't keep. And so that's really, to me, what keeps me from saying what's happening before it has happened and we have actually shipped the product. The other thing I'll say is often when people do know who we're talking to and I ask for questions, the questions we receive are so in the weeds that I know the CEOs won't know the answers.
And that's, to me, the biggest thing I've learned doing the coder all this time is the CEOs are way up in the clouds. And everyone wants to know why the button on the product doesn't work the way they want the button to work. And the CEOs sometimes don't even know the button is there. I think that disconnect is sometimes really, really useful, but I think fundamentally unsatisfying.
I'm always on edge with that.
One thing I thought of when I was reading this question was also the fact that sometimes with interviews, you want the guests to not know what you're going to ask them, right? I could imagine there's PR people who look at your Blue Sky feed, look at your Threads feed, and they go, oh, look, he's soliciting questions.
We've got to get him a full list of these right away so he knows how to prepare for these. I guess there's probably an element of that going on, too.
Yeah, I mean... One of the things about Decoder is it's part of The Verge. The Verge is annoyingly journalistic, and we do not allow anyone control over what questions get asked. We don't give the questions away beforehand. There's no guardrails. That is not the case on a bunch of other podcasts.
I will let you imagine what other podcasts this might be, but we are, you got to show up and you got to take the heat and, you know, soliciting a bunch of questions is fun, but it's also, it takes the edge off because they're prepared for a bunch of the questions. And I would like people to be more candid. That's not to say we won't do it. And sometimes we'll find ways to do it.
There are guests that I think we would schedule and we would know they would show up on time and like, We would around events, particularly or news moments that tends to happen on schedule. But I want to make sure that we produce the most candid show we can and giving away a bunch of questions kind of cuts against that.
Listener named Brian wrote in about our April episode where we talked to Verge reporter Liz Lopato about the rise and fall of Vice Media, which is honestly very fun. I love any time we can talk to Liz. Brian wrote, the conversation about funding sources and problems at Vice got me thinking, how can we as readers and listeners best support what you do at the Verge?
I mean, when you say I segued into this, every question segues in to pay us $7. The best way to support us right now is to obviously sign up for our subscription, which is $7 a month or $50 a year. If you pay the annual and you live in the United States, we'll ship you a book, which is really cool. It's called Content Goblins. We're working out how to do it internationally. It's complicated.
But that is the best way to support us right now. And I'm happy to say the subscription has been a couple weeks by the time you listen to this. It's blowing away our goals. We're very pleased with it. That's the best way. If you're an Apple News Plus subscriber and you don't want to pay us directly, you can reach us at Apple News Plus. That also pays us money.
There's all the other stuff we do, right? There's advertising and affiliate links. But truly the best way to support us is to directly support us. And that is going to give us the cushion we need to make sure we can still do rigorous independent journalism without having to do what every other kind of media creator is doing, which is go and take brand deals.
And I really just don't want to do that anymore. I'm allergic to it. I don't think I can. And you look at the rest of the media ecosystem, you look at creators and influencers, they're all doing that because the platforms aren't paying them high enough rates. So they have to take brand deals to make enough money to be sustainable. We can't do that. We wouldn't do that.
I think that's incompatible with journalism fundamentally. So we're asking for direct support from our readers and listeners. That's the answer.
One thing we've been asked a lot in the last couple of weeks since the subscription launched is, of course, ad-free decoder. Is there a tier of the subscription that is going to include that? Is there something that people can do to remove the ads on decoder? Is that even financially feasible? How does that work?
We get this question a lot from Vertcast listeners as well. I would point people to the episode of the Vertcast where our publisher, Helen Avlak, talked about it in detail and why it's difficult to do ad-free podcasts. You don't want to take half of your listeners out of the various algorithmic discovery platforms.
There's some complication there, but we can see that people want it, and we just have to figure out pricing to make it work and be sustainable and also affordable.
We've also had some listeners write in asking about ways to get around the paywall, not to, you know, stiff us, but for students or academic purposes like gift links or organizational memberships. Milton, who is a professor of digital marketing at University of Wisconsin at Madison said, asked us, how do I buy more?
I'd love to figure out a way to give my students easy access to articles so they don't encounter the paywall. I'd be willing to help offset the cost of individual articles I leverage. For example, could we partner together on an affiliate link approach? I'd also turn it into a meta assignment for them to learn more about affiliate marketing and the disruption in the creator and content economy.
Could we let this professor buy more Verge?
I feel like if I learned anything from doing Decoder all this time, it's we should definitely find ways to take the money. That's a very good question. Our paywall right now is designed to just convert the first set of people we think are most likely to pay. And all these other business models you're describing are really smart. And I think we have to get there in the product.
But the first step was just launching it and seeing how people reacted to it. And now we'll iterate from there. We have a lot of ideas for other kinds of businesses we can run. Certainly our staff would love gift links. They ask me every day for gift links. There are lots of readers out there who would love to pay to just open things up for an amount of time.
There are advertisers on the publishing side, my colleague Helen Havlack, who runs that side of the business. She has advertisers who want to just make the site free, like do windowing on the site with the paywall, which is really interesting, not a business we've had before. So there's all these opportunities.
We needed to just launch the thing, and now we need to start going down the list one by one.
We have to take a quick break. We'll be back in just a minute.
Support for Decoder comes from Stripe. Your customers aren't thinking about your payment management software as long as it works. But the moment your checkout now button stops functioning, customers will start thinking about it a lot. Or they'll go somewhere that can actually accept their payment. Stripe can help you make sure that doesn't happen.
Stripe handles the complexity of financial infrastructure, helping you offer a seamless experience for business owners and their customers. Stripe makes sure that your customers see their local currency and preferred payment methods when they shop. So checking out never feels like a chore.
Stripe offers a suite of specialized features and tools to power businesses of all sizes, like Stripe Billing, which makes it easy to handle subscription-based charges, invoicing, and recurring revenue management needs. Learn how Stripe helps companies of all sizes make progress at stripe.com. That's stripe.com to learn more. Stripe. Make progress.
We're back with Nilay and the Decoder team, answering your biggest questions and talking about what's going to happen in 2025. We're going to change gears a little bit here and start asking some bigger questions, less process, more about your thoughts on platforms and technology. So one question we have from Brian Penny on Blue Sky.
The Verge was one of the more prominent media outlets on threads, but for the past month, it's clear you're all in on Blue Sky much more. What caused this change? And it can't be just that meta deprioritizes news because that was true for the past year and you were all still using it as your main platform anyway. So what's going on with Blue Sky?
You know, it's interesting is the day you're asking this question, I believe our audience team is using the Threads API to put even more content on Threads. And that is kind of the underlying issue is these platforms are new and they did not have the big suite of publishing tools that you need for a newsroom of our scale that Twitter had or Facebook once had in the past.
so everything was manual so if you saw us turning our attention literally we would turn the attention from one platform to another because we just have people doing it manually that is changing we are generally very interested in decentralized social media and decentralized distribution because ultimately what that allows us to do is control our own distribution which is a real decoder theme
In a real Verge theme, we want to be in control of our own distribution. We want the Verge to be independent of tech companies and their platforms because we have to report on them. So yes, there's a little bit of Meta owns threads. We don't want to be all in on more Meta distribution.
Everyone here has worked at a media company that has been destroyed by Meta's promises of distribution that have gone away. That's just a reality of working in media and the internet over the past 20 years, but But it's not that we're against or for any platform.
It's literally Blue Sky was taking off and our audience team turned its attention to manually posting on a platform that had the heat for a minute. And then in the meantime, the API-based publishing on threads, those tools started to get built. And that is being automated today. So you're going to see us engage in a lot of these platforms.
And where I really want to get to is lighting up Federation on our own site. So we're natively distributing content to these platforms and collecting the engagement data in return, which I think is the promise of interoperability. I have no idea how it's going to work. I'll let you know over the course of the next year.
Speaking of Blue Sky, that's where this next question came from. And I'm already laughing. It's from Caridot Peters, who writes, there's too much talk about podcasts as the primary source of information as news. And I get why. No or fewer paywalls. Do you see any alternative? Do you, Neil, I still believe in websites. Oh my God.
Also, with so many newsletters in addition to news sites, could an RSS-type solution be in the future?
An RSS solution to newsletters? to information. To information. I am still a writer. I think it is like a tragic quirk of fate that I've ended up making more podcasts than anything else. My heart is in writing and reading. I think we do the transcript of Decoder every week just because it's easier for me to read it, and I prefer reading interviews to listening to them. That's all very funny.
Like, there's just something personally comedic about that, that I've ended up producing more audio than text, and then I read my own text. I think there's a lot of benefit in reading, in text. A long time ago, when the whole media was pivoting to video, I was like, we still need to write things down. Making a video, making a podcast is a lot of reading and writing.
Kate and Nick do a lot of reading and writing to prep for every single interview that we do. It is not easy to listen to 5,000 podcasts to prep for the next podcast. You need the text. I know our listeners, I think, are constantly telling us they appreciate the transcripts. It makes them reference material in a really important way. So we're going to keep doing that.
I think there's a great future for writing. I think there's a great future for websites. I understand that there's a big audience that only expects everything to be a three-hour YouTube podcast. Pendulums swing back and forth, and eventually I think that audience might get jobs. They might value their time differently.
So aren't we still the last website on Earth? That's the verge, right?
Yeah, we're very committed to running our website. I like not working for any of these platforms. And again, I have a lot of sympathy for people in the creator economy. They've built bigger businesses than I have in a lot of cases. But we are not subject to any. I never have to watch an Adam Mosseri video on Instagram telling me how the algorithm is changing or what videos to make and do anything.
And a lot of Instagram creators watch those videos, and they change their strategies. That's weird. And we're so ferociously independent that having our own website that people come to directly just insulates us from all of that. And so I think we're going to, as long as there's a web, we're going to have a website.
It's funny you say all that because kind of a cornerstone of the decoder research process is going to a search engine, typing in some information, and finding an article from 20 years ago that tells you the date of when something happened or, you know, a merger from 2002. And that's really, really important to have that information available to you.
And it does feel like that is harder and harder to find the further out you go. Which ties into our question from Ty on Blue Sky, who said, with instability increasing in a number of areas, will a focus on permanence see a renewed push? All our digital things are only as good as stability allows.
I think this is just my lawyer background coming into play. but I think there's gonna be a market for paid search products. And specifically what I mean is if you're an attorney, you pay a lot of money for LexisNexis or Westlaw, which are databases of legal information and articles that are there, that are organized to be searchable, that have ancillary information around those searches.
And it's all designed because there's a market for attorneys and paralegals doing very specialized search And then you can charge them really high rates to guarantee good results. I'm not sure that's totally one-to-one. You can't just do that for consumers.
But you can see for a whole bunch of other kinds of markets, you're going to be able to pay for, hey, I don't want AI summaries of garbage here. I just need the answer. I just need some more specialized search. And there was a move to that. They called it vertical search analytics. Back in the day, you had your kayaks and your Yelps and Google kind of ate them.
And Google's in so much trouble right now with various regulators that it feels like maybe the vertical search providers will be able to do something else. But that thing you're describing where I just need to know this is real information and it's organized and it's going to deliver me those routes quickly, because our general search providers are so messy and they're trying to do so many things,
It feels like someone should be able to show up and make a cheaper Westlaw for consumers or journalists or other researchers who just need the information. I don't know what those markets look like, but you can just see the demand for it.
A lot, a lot, a lot of listeners wrote in about our interview this fall with Intuit CEO Sasan Godarzi. One example is from Jamal Khan, who wrote, "...I appreciated the tax reform questions you asked the Intuit CEO in your recent interview. They were fair, and it would have been a glaring omission if you hadn't brought it up. However, early episodes of the podcast were missing that edge."
He adds, that's why I think now is a good time to revisit some of those early interviews. Back then, it felt like you were inviting guests because you thought their companies were doing something cool and you wanted to share that with your audience. Some of those companies failed spectacularly in realizing the vision they sold us.
And so the question is, can we go back to some of our guests from the first year or two of the show and ask them newer, harder questions? Jamal specifically called out UiPath CEO Daniel Dines as someone he wants to hear from again.
It's funny you mention that. UiPath literally reached out 20 minutes before we started recording and offered us Daniel again. Daniel stopped being the CEO, went to be the chief innovation officer, is now the CEO again. That's just pure decoder bait. We'll have him back just to talk about that. What happened there? But UiPath is fascinating, right?
They were already automating a bunch of systems, and now you get agentic AI. That might be a new kind of innovation that totally undoes everything they were doing. We'll have that conversation again. Yeah. You know, the Intuit one is a fascinating case, and Nick and Kate, you should talk about it too, because we just asked that question at the end, and it was fine.
Like, I think most people, the reason I was, like, we just run this at the top unedited is because, yeah, it was a little spicy, and he was a little unhappy with me, but that happens in every episode. What made it, like, controversial is they asked us to take it down. Yeah.
If I can get to a place in every episode where the company asks me to take it down and I can run X company asks me to delete this episode, I would do it every week. They don't usually screw up that bad. And I think all of us were deeply surprised by that.
It was the most unhinged professional comms email I have gotten since I became a full-time journalist in 2012. And like, I spent the first half of my career at Consumerist ticking off companies every week.
Yeah. I'll give people an example. We had Philips on the show. We talked a bunch about AI and healthcare, Philips is a healthcare company, and the core issue there was they had shipped a bunch of CPAP machines and ventilators that had some foam that was degrading. There was a recall. Some people had gotten sick, allegedly. That was the same style of questioning in my mind. Here's the problem.
Talk about the problem. Here's the controversy. I'm looking at these lawsuits. Fine. They just didn't ask us to delete it. The CEO knew the questions were coming because obviously those questions are going to come. He had his answers. Maybe you were unhappy with the answers. Maybe the answers are boilerplate. Whatever it is.
but they weren't so unhappy with their own stock answers that they asked us to pull the episode down. So we do try to get to the controversy in every episode. It's just very rare that a company makes its own non-answer controversial in that specific way. I hope Sasan comes back. I'll talk to him again. I suspect he will not. But we're going to keep trying to ask the questions as hard as we can.
And I agree that we had some companies on in the beginning of the show's run that we should have on again because they made a bunch of promises, made a bunch of claims that maybe came true or did not come true. And I think why didn't this work is such an interesting thing to talk about on Decoder. And I know a bunch of those guests would actually be happy to explain.
why their ideas didn't work or why their plans didn't work and how they pivoted, because that's ultimately the essence of the jobs they're doing.
Well, listener Mike Espinosa had thoughts about what we could ask guests. He says he seems to remember, Nila, you asking a bunch of people, when do you read email? He misses hearing that question. He thinks it would be especially good with repeat guests. He also gives his personal email philosophy. So he wants to know, what is your email philosophy? I'm curious, what is your email philosophy?
Are you an inbox zero person? Are you reading your email all day long? More importantly, should we start asking guests this question again? Do you think we'd get interesting answers?
Yeah. Well, I'm not. What's Mike's email philosophy? Give me the summary.
So Mike's email philosophy is he doesn't, he says, my answer for the last few years is he doesn't read his email. He filters anything that was sent to me outside of his org and not addressed to me. Even CCs are filtered.
For the most part, nobody notices, and when they do, I just say the truth, which is that I get so much email that sometimes I look at something, and then I forget to respond to it, and it gets buried. Nobody ever questions him on that, and nobody stays mad, he says.
He archives everything, so if it does come up, he can search for it and pretend like he read it, which this honestly sounds like a great email philosophy.
I feel like I need to get a text expander for, sorry, I missed this, and it would just be fun. I agree, no one cares about that anymore. My philosophy is largely the same, although there's an element of things I just can't miss, especially when you're trying to book a show with CEOs. I just need to make sure I see those emails.
So I do a lot of scanning in the morning and the afternoon to just make sure I see everything. The responding happens later. I... I don't know if the email is a solvable problem. I think I'd probably need to sign up for a service like SaneBox or whatever David Pierce is talking about all the time to solve my inbox problems.
The reason I started asking executives when they check their email is when we were prototyping Decoder as episodes of the Verge cast, which we did for a while, I was just trying to figure out how to structure the show. And what I mean by that is if you sign up to interview somebody every week forever, you now have a forever project. And forever projects are bad. I'm incompatible with them.
I want things to have beginnings and ends. I want to know when something has run its course and it's time to change. And so to me, the forever project of be interesting with a new CEO every week, I found it to be pretty challenging. And I was like, what is the way to structure these interviews so they deliver enough content? consistent good stuff to people.
And so I was just trying out different ones. I asked people how they structured their time for a long time. I definitely asked, when do you do email? That was part of how you structure your time. And I settled on, how do you make decisions? Because I figured everyone would have to have an answer to that question. You ask people how they structure their time.
Some CEOs just look blankly at their assistants. they do not know. And so like, I wanted to have some question that everybody had to have an answer for that would be revealing, right? Like that, how do you make decisions is in its way, a very low stakes question in its way, also the highest stakes question because
Because if you're the CEO of a company and you explain how you make decisions, a bunch of people will have heard how you make decisions. And they will try to use that framework to get their decisions made or to convince you of something. So it felt like the right mix of very low stakes. Everybody has to answer this question. It's not controversial. You're not going to go to jail.
The European regulators are not going to show up on your doorstep because you answered how you make decisions. But it's high stakes inside your organization. And that very quickly led to how have you structured the company, which is now the other decoder question, which is very organic.
Because it turns out how do you make decisions and how is your company structured are kind of the same question, just expressed in two different ways. So maybe we'll go back to some of these more practical. How do you do your email? When do you send meeting agendas out? Like all this stuff, like the sort of tactical elements of just like having a job I think are interesting to people.
But that stuff changes so fast. that I don't wanna get lost in it. Because the cultural norm of just telling people you didn't read their email is brand new, right? That was not true, and it's still not true at some companies. So I'm always on the market for more questions that provide structure to the show and allow me to consistently deliver it and make the show feel less like a forever project.
I can just print 100 answers to here's how I make decisions into a book, and we can call that a book. I don't think that would be a very good book, but you could do it, and that provides me a measure of peace. That like, oh, this project could come to an end and I would have something to show for it. And I'm always in the market for more questions that do that. But then there's the news.
And I want to make sure we stay in the balance of the big evergreen show that has the philosophical questions. And like, hey, are you doing good stuff as far as your products work? Because that seems important too.
Our work from home episode this fall also generated a lot of feedback from listeners. I think it was probably tied with Intuit, actually, for the most feedback we got this year. One email from a listener named Rohit Kabra really captured the sentiment.
He wrote, currently, I'm the founder of a growing startup and the work from home versus return to office debate is one I have frequently with other founders. He wrote that he favors a hybrid approach for his team and was excited to hear our take on it. But then he didn't like our take on it. He said our perspective felt a bit narrow and even dismissive by focusing primarily on a subset of society.
And he listed his concerns about the impact on interns and new workers who don't learn how to be in a workplace very well without one. He wrote about leadership quality, which is that maybe poor managers just really do do better with people in the office and you can't expect every manager to be great.
He wrote about cities and office culture, mental health, isolation, and the growing divide between knowledge sector work you can do with a laptop from anywhere and every other kind of job in the world that requires you to be hands-on. And the questions from this are, where do you fall on the remote versus in-office debate? And how does the decoder team operate?
Well, that one's easy. The decoder team is fully remote. I've never even seen these people. No, it's not true. We see each other a couple times a year, but we are fully remote. And I don't know that we could make this show if not for the transition to being fully remote.
like our guests are never with us it's a nice surprise and an added bonus when the guests are in the office with me so if we weren't able to just produce the show remotely i don't know we'd be able to do two episodes a week i don't know that we'd be able to get all the ceos that we get it's just the reality of these kinds of shows right now and then we're all be able there are like there are children on our team like everyone's able to take care of their families like that's the thing you need to do when the
recording schedule of CEOs showing up whenever they can show up is as chaotic as it is like we all just need to be able to work around that thing and that sort of I don't know Kate and Nick you tell me but that sort of in my mind means it's better that we're remote like if we're all in the office together all the time I think there are some conversations that would be a lot easier but I think we would just be in the office late at night quite often
A huge number of the CEOs we talk to are in California or Europe. Nilay is based in New York City. I'm based in D.C. And yeah, we could not go to New York or Europe or San Francisco for every week we talk to all these people.
Yeah, so that's just the decoder answer. How should you make your company? Should you all be back in the office? Should you not be back in the office? Should you be hybrid? The Verge is mostly hybrid. There's a bunch of people who go to the office in New York quite often. I'm in New York. I try to go a couple times a week. Lately, I've been very bad because it's the holidays.
And quite frankly, I like working from home. I'm more productive at home. I agree that it's much harder to manage remotely. I think it is staggeringly hard, especially for new managers. Telling people what to do is really, really hard. Giving people negative feedback, incredibly hard.
Doing it on Slack or over Zoom and then you hang up and you're still just alone in the room that you were in when you got the negative feedback is, I think, one of the worst experiences you can have.
Someone just said something to me about my performance that wasn't good or gave me some negative feedback, and then I hung up the Zoom call, and now I'm just alone in this room, and I have to do the next thing and pretend that that didn't happen to me. Just as an emotional experience, that's bad receiving. It's bad giving that experience to people. I agree.
I don't know how to solve that problem without reorganizing the American economy. Which if you would allow me to do, I would do. That appears to be how the American economy will be organized, just like one person's whims. I could be that guy. But I think this debate is going to swing back and forth.
And what I know from that episode is 50% of people feel strongly one way and 50% of people feel strongly the other way. And our episode tried to chart that middle ground. And I am confident that irritated everybody. But that is the middle ground most companies are trying to chart.
We have to take a short break. We'll be right back.
Support for Decoder comes from Stripe. Imagine a world with 99% reliability. You can glance at your weather app and know exactly what to wear. Your flights would always leave on time. Your flaky friend would actually bring a main dish to the potluck. Sadly, that's not the timeline we're living in. That's what makes Stripe's payment platform so remarkable.
When they say they offer 99.999% reliability, they actually mean it. Stripe is a global payments leader serving millions of businesses of all sizes, including Hertz, OpenAI, and DoorDash. And their industry-leading reliability is a big reason why. Historically, Stripe has managed to maintain nearly flawless uptime, even on Black Friday and Cyber Monday.
Stripe makes it easy for businesses to accept popular payment methods around the world. The platform offers advanced billing software to help you grow your revenue and work seamlessly with virtually any business model. Learn how Stripe helps companies of all sizes make progress at Stripe.com. That's Stripe.com to learn more. Stripe. Make progress.
We're back with Nilay and the Decoder team answering your biggest questions and looking ahead to 2025. So a lot of listeners have written in with, of course, their thoughts about AI, the coder's most explored theme of 2024. John Pickerton wrote saying that he would love a dedicated podcast or miniseries on something like the intersection of AI and software engineering.
He was a big fan of the GitHub episode. He says there is a lot going on in the space, including lots of people himself exploring running models locally on their own GPUs. Cam wrote in that there is a topic he would love to hear more about on the Decoder podcast, and that's AI in the defense sector. He says he doesn't hear a lot about that.
He would love if a podcast explored a lot of the questions around AI and the defense sector, like investments, if the tech is working, who's benefiting from it, the ethical questions around it. But the question to you, Nilay, is what do you think are the most unexplored AI topics right now? And what are you personally interested in exploring about AI more in 2025?
Well, let's start with defense, because I think defense is super interesting. We haven't done a lot on it. Historically, The Verge doesn't do war coverage. We're sort of resolutely an anti-war publication. And we've had long editorial conversations for 13 years about, should we glorify drones or cool fighter jets? Because they're tools of war, and we're resolutely an anti-war publication.
That said... The thing that will happen in the Trump administration that is utterly fascinating to me is the massive split between the Defense Department as it exists today and the Defense Department as envisioned by Palmer Luckey and Andrew Hill, by Peter Thiel at Palantir, by Marc Andreessen who's making huge investments here.
There's a reason Elon Musk is tweeting once a week that the F-35 is stupid. It's because he wants to sell AI-enabled drones. There is an incredible profile in Tablet Magazine of Palmer Luckey called American Vulcan, which you should just go read. I'm not saying I agree with it. But in it, he lays out what is essentially a vision for the AI-enabled defense of America and Israel.
Where I will put swarms of AI drones into the sky to protect these countries from outside aggressors. And he describes himself as part of the warrior class. That's all coming. That's a big split. We should stop having aircraft carriers as a thing the tech industry is saying because they want AI-enabled drones in the sky. Yeah, we should probably cover that.
Like, I don't think the Pentagon and its supplier base and the massive number of workers in this industry are just going to accept that level of change without a fight. So we should probably talk about that. One of the things that's driving all that is these VCs are sort of out of giant markets to conquer. in the defense industry is a pretty big one. So it's coming.
We should find a way to cover that. I don't know how to do that without betraying my own personal bias of being the editor-in-chief of a resolutely anti-war publication. But it's coming, and I think it's worth exploring because it's happening really, really fast. The other unexplored theme of AI that I'm really interested in is whether it is actually the enabling technology people want it to be.
And I think you've heard me ask a bunch of CEOs this question already, and we need to just focus on it a little bit more and put it into relief that. If the models don't get better, if the scaling laws don't keep going and the underlying models don't get more capable, then there will be a limit to the products we can build with them.
And you have to invent some other stuff to make those products do what they want consistently. And that to me, it feels like we're just beginning to peel back saying whether or not that's true. And I think over the next year, either this bubble pops or it doesn't. And a poor question for me is like can the technology do what people want it to do?
If it can do what people want it to do, yeah, I think how we program computers generally will change. Natural language programming will just be a thing and that will be a strange reality because we're used to very deterministic computers and now having natural language computers – wacky.
But I think this year is the year where we find out if the underlying technology of LLMs can do all the things that people are dreaming of in the products they want to build.
because our interview with Rabbit CEO Jesse Liu also generated listener feedback. First time listener John wrote in to say he tried the Rabbit R1 episode and was not sure what to make the most of. He said the back and forth about scraping data and getting blocked by big companies was fascinating.
He said they're doing something super brittle, so it was really satisfying seeing the interview drill down on that. He adds that by the end, he had been won over to the CEO's way of thinking a little bit. They could still be crushed at any time. And he went on with basically the question, where do you think Rabbit goes from here?
Is this kind of thing going to fade away or are we just way too early for AI hardware?
Oh, boy. I think those are two different questions. It seems clear that whatever next version of phone there is has some level of agentic AI in it. Even the next version of Siri that Apple's promising to ship sometime next year with app intents where you can speak in natural language to Siri and the apps on your phone will respond to that and that will all be more natural language.
That's the dream of Siri. It has been for a long time. So I think that's coming. Whether or not it works, whether or not Apple intelligence on my phone works today as well as anyone wants it to is an open question. Whether that works as well as anyone wants it to is an open question. But that's coming to the phones.
Google is already shipping a bunch of that with Google Assistant on their phones today. I think that leaves not a lot of space for AI hardware. And what I think the AI hardware boom that we saw last year was really all about was here's a new interface paradigm, right? ChatGPD lets you do natural language interface.
We can build another piece of hardware around that interface paradigm the same way that we built iPods around click wheels and smartphones around touchscreens. And maybe this is the one that lets us build a new interface. app ecosystem and market for applications. And everyone got really excited about that. And then, like I keep saying, the LLMs weren't good enough to deliver the product.
If the Humane pin had actually worked and had low latency and you could just talk to it and it just went off and did stuff, I think we'd be talking about it as a competitor to the phone. If the Rabbit actually worked, if you could just talk to it and it could go out and do stuff, I think we would be talking about it as a potential competitor to the phone.
The problem, and I think this really came out in the interview with Jesse, is they are in the fake it till you make it phase. They were doing pretty basic robotic process automation of the Spotify website until the interview, basically, when they were rolling out their more agentic system that could intelligently navigate the website.
But at the end of the day, you still have a Spotify website in the background. You still have the DoorDash website in the background. And those companies, I think, are not just going to hand their customers over. They won't hand them over to Apple. They won't hand them over to Google without payments until that economy shakes out.
Yeah, what you've got is a user interface that isn't good enough and a business model that is definitely not good enough. And I think that's just going to keep the hardware – the AI hardware market pretty constrained while Apple and Google build up the business models and then they just ship the next version of the phone. So I'm pretty –
pessimistic about all of it, but I see why people were excited because when you have the new user interface tool, when you have a click wheel or a multi-touch screen, it is natural to think that you will have a paradigm shift in the hardware and natural language is the user interface that people have dreamed about the entire time.
So Jordan from blue sky gave us the perfect wrap up question for the Q and a portion of this, which is Nilay. What do you think might be the overarching theme for tech in 2025?
Look, I'm an optimistic person. I love technology. I'm just going to say this answer. It's bubbles popping. Left and right. And there's two in particular that I will just point to. One, I think the capability level of the AI tools is going to lead to some amount of bubble popping. Just the idea that you can throw AI into a product and that justifies some amount of hype or whatever.
More people are going to have more experience with the tools. They're going to understand what they can and can't do. More people are going to have experienced... the sort of asinine Apple intelligence summaries of their notifications on their iPhones. And the bloom is just going to come off the rose. And then maybe we'll build some real products.
This is usually what happens when a tech bubble pops, right? There's the trough of disillusionment. We build some real products and maybe the valuation comes back up. But that one's coming. It might already be here. Like in some real way, it might already be here. When our social networks get choked to death with weird AI videos, something's going to happen there.
Like some consumer demand is going to move. And that might already be happening. So that's one that I see pretty clearly. We're going to spend some time investigating. The other one that I see is somewhat related to this, which is that I think the influencer media economy is sort of at a peak. I'm not saying it's all the way at the peak. I'm not saying it's going to crash.
I just – once you get a president, there's kind of nowhere else to go, right? Like your influence is kind of at the highest it can be. And the influencer media economy created an election cycle and that – That's fascinating. None of it is economically sustainable, right? This is the thing I'm saying. The platforms don't pay high enough rates and there are more creators every day.
YouTube is not growing its revenue as fast as it's growing its creator base. So now you're just doing division. right? There's some amount of money coming in. There's some amount of creators to share that money with. There are more creators than there is revenue. The rates are going down. And so you see the creators are having to turn to brand deals.
You see the push is for everyone to make more content all the time. You see the platforms telling creators to use AI tools to make more content, to respond to comments using AI avatars so they can focus on making more videos. Like,
All of that push towards more, more, more, more, more is really just about the imbalance between the number of creators coming onto the platforms and the amount of revenue coming to the platforms. And the rates are going down. And I think that is – when you see something that looks unsustainable on paper, on the spreadsheet, it probably is unsustainable. And something else will happen there.
I'm not sure exactly what it is. But that's the one where if you ask me which of those two I'm more convinced will change this year, it's influencers, not AI. I think everyone is predicting AI will come to some sort of head. The scaling laws will run out. The models will be less capable. Whatever it is. Yeah, I buy that. I think people are going to have experiences with the products.
And the truth in products always comes out. You can't run from whether or not the thing works. The influencer bubble one is like, I, no one can run from math. Like no one can run from the fact that being a YouTuber eventually burns you out. And that's just the reality of, of being a YouTuber, Instagram creator.
And until that's solved, I, I, I think that economy is gonna start to turn in a different way. And that will be really interesting. It's gonna be one of the projects that we cover at the verge, all of 2025, we're already talking about how many stories there are to do in that zone. And we'll probably have a lot of it on decoder.
So speaking of 2025, we here at Decoder are already very deep into planning our 2025 out. Since I am the person who has to be in charge of logistics, I can say we have booked interviews through almost the end of March already, and it's not quite Christmas. We can't talk about most of those guests yet for a lot of good reasons, but we have some big ambitions.
Nilay, who do you think we should try to get on Decoder in 2025?
I think we should absolutely try to get Sam Altman on the show. I think he'll do it. We just have to keep pushing on it. Let him know. If you're a Dakota listener, let him know that you'd like him to be on the show. It always helps. I really want to try to get Bob Iger. I think the media is in just a weird space right now, and I don't know that anyone knows how Bob Iger is making decisions.
And I think he's also good. I think Bob Iger likes talking about how he runs Disney. So I think he's very gettable. You know the really interesting one? Correct me if I'm wrong. I don't think anyone has ever asked us to get Tim Cook. Not once. I have never asked to interview. It has never even occurred to me. And I think it's because I've never seen an interesting interview with Tim Cook.
So I'm not sure if that's the challenge I should put on our plate. But one of the things I want to do with the show next year is make sure all of the interviews are interesting. Let all the people show up with something to say. The split you can see already on the show is we joke about it all the time on our team. Our founder interviews are always utterly fascinating.
Like Brian Chesky shows up and he's just raring to go because he founded Airbnb. He's confident in the thing he made. Like our founders are almost universally more interesting than like the McKinsey robots. And sometimes companies are just run by McKinsey robots.
And so I want to make sure we head towards people with something to say and we are a little more discerning and make sure people show up without just consultant speak. So we'll work on that. But yeah, there's some people on our list I think are going to be pretty interesting.
We're also heading into almost the one-year anniversary of our second episode. Eli, what do you think we should be covering in the second episode, and why are you excited about that?
The second episode, I think, is going to be in a really challenging spot next year. Decoder as a whole is sort of a system show. Like, here's how companies work, and then you have all these people who show up, and they talk about how their companies work. And most companies are kind of the same. You have a product. You try to make the product more efficiently.
You try to make sure that people are happy. You try to pivot the product or iterate the product. They're mostly the same. No matter what you're making, you have some inputs. You try to refine them into outputs. You try to sell the outputs for more than you bought the inputs. It's a company.
And then we have the regulatory piece where we're like, here's how the law works that constrains these companies. And next year is just all up for grabs because of the Trump administration. And so the idea that we have a second episode that will explain copyright law to people...
when you have a bunch of Trump administration officials who are just like, here's some ideas I have about copyright law, they're related to nothing, is going to be complicated. Section 230, we have done infinity explainers about Section 230. Kate, you've probably written infinity explainers about Section 230 in your career at Protocol and other places.
That's Brendan Carr is going to be the chair of the FCC, and his stated plan is to issue a ruling from the FCC that reinterprets Section 230 for the courts. Just like stepping through all of the ways that is nonsensical is an episode of Decoder. I can do it right now very quickly. One, the FCC can't do that. Two, courts don't listen to agencies.
Three, there was just a Supreme Court ruling that said courts definitely don't have to listen to agencies. Like, it's a weird cluster that I think is going to make the second episode both very challenging. but also very important. And by challenging, I mean, I think some of our episodes are going to be fundamentally unsatisfying. We will not explain anything.
What we will explain is that the explanation that we used to have is now out the window. And I think that is a service and we should do that service, but it's going to be a wild ride.
All right. I think that's about it.
Wait, wait, wait, wait. Hold on. I want to ask you two a question. What do you think we should do? You are the two who have to deal with me. So what should we do with the show?
I'm really interested in exploring the topic of what the introduction of generative AI has done to education. I have two kids. One is in middle school. One is in first grade. And they are glued to their iPads at all times. And I am trying to teach them what search is. And it's real complicated. And so I'm hoping we can get some teachers on to talk about what the deal is for them.
I would love to do more about robots. I think robots are a fascinating industry and I think they're going through a lot of changes right now. I just read a fascinating profile actually in the New Yorker about generative AI tech being applied to robotics in a way that is fascinating. Achieving some interesting breakthroughs.
I think there's going to be a lot of really fascinating stories happening at the intersection of AI and then the things that we want AI to do in the real world, which are historically very constrained by how bad robots are at maneuvering the real world. But that could change. And I think there's a lot of fun stories to be had there.
A lot of really challenging stories, too, about what a world of robots will do and what that means.
Yeah. You know, it's funny. I can mash those two together, which is my niece and nephew are in college, education, and food delivery robots are just a thing. It's just a thing a bunch of college students are used to on campuses across the country. And it's like, those aren't the robots anybody thought we were getting. but they're the robots everyone is used to.
They're just an accepted part of campus life. And there's something there that I feel like is worth an hour of our time. This is weird, right? And there's cultural norms around these robots that are just rolling around campuses. I think it's utterly fascinating.
And that's it for our 2024 end of the year wrap up. I'd like to thank Nilay for being the Decoder guest for a change and Kate for joining me on the show. And thank you for listening. I hope you enjoyed it. We had a lot of fun and we're going to do this again next year.
If you have thoughts about this episode, what you'd like to hear more of or answers to those questions we discussed at the end, like who should we have on the show in 2025 and what our second episode should focus on, please do email us at decoderattheverge.com. We hope this episode proved that we really do read every email. You can also hit up Nilay directly on Threads or Blue Sky.
He's reckless1280. We also have a TikTok. Check it out at DecoderPod. It's a lot of fun. If you like Decoder, please share it with your friends and subscribe wherever you get your podcasts. And if you really love the show, give us that five-star review. Decoder is a production of The Verge and is part of the Vox Media Podcast Network.
Our producers are Kate Cox and myself, Nick Statt, and our editor is Callie Wright. Our supervising producer is Liam James. The Decoder music is by Breakmaster Cylinder. See you next time.
Support for this episode of Decoder came from Stripe. Stripe builds financial infrastructure that millions of ambitious companies use to help support their biggest, boldest ideas. And the platform has something to offer companies of all sizes, from brand new startups to established global businesses.
Stripe has processed trillions of transactions, and the platform is constantly analyzing past sales to generate revenue-growing insights for its partners. Its suite of optimized checkout features can help boost revenue by intelligently converting prices to local currencies, offering customer-specific payment methods based on past transactions, and more.
Learn how Stripe helps companies of all sizes make progress at stripe.com. That's stripe.com to learn more. Stripe. Make progress.