The Times Sues OpenAI + A Debate Over iMessage + Our New Year’s Tech Resolutions

“This is sort of the big kahuna.”

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

• 0:00Speaker 00

  AWS AI is how the world's leading organizations are transforming their industries, not in theory, but in production, at scale, right now. Like turning race day data into real-time insights for Formula One fans, reinventing the journey for every United traveler, and personalizing every trip with booking.com. However big the question for your business or industry, AWS AI is how.

• 0:26Kevin Roose?

  I got the call about the lawsuit at the funniest possible time. I was on vacation and I was at a bird sanctuary.

• 0:36Kevin Roose?

  What were you doing at a bird sanctuary? You know, they have these places where you can like go see like parrots and you know, toucans and- Yeah, aren't they called zoos? This is like a small, specific sanctuary for wounded birds. Wait, and they're all wounded too? Well, some of them are wounded, yes. So they bring them in, they rehabilitate them, they give you these little cups of seeds, and you like hold the cups, and then the birds come and land on you and eat the seeds out of your cup. And was that how you got bird flu this holiday season? Yes, so I'm walking around, I have like two parrots and like another bird like on me. I'm sitting there holding this cup and I like looked down at my watch and it's like a notification that's like, please call me, the New York Times is suing OpenAI.

• 1:25Kevin Roose?

  Oh, boy. Oh, my gosh.

• 1:33Kevin Roose?

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

• 1:35Casey Newton?

  I'm Casey Newton from Platformer.

• 1:36Kevin Roose?

  And this is Hard Fork.

• 1:37Casey Newton?

  This week, The New York Times is suing OpenAI. We'll tell you what's at stake. Then, Beeper CEO Eric Michakovsky joins us to talk about how his company hacked iMessage so that Android users' green bubbles briefly and gloriously turned blue. And finally, Kevin and I trade our New Year's tech resolutions.

• 2:09Kevin Roose?

  How was your break, by the way?

• 2:11Casey Newton?

  A great break. Got to see so many friends and family, bring in the new year in style, and develop that sort of divine sense of chill that you really only can get if you're able to take like two sustained weeks of vacation. And then, you know, got on a plane and I would say that spirit was completely dashed.

• 2:31Kevin Roose?

  What happened?

• 2:32Casey Newton?

  So I flew out of the Burbank airport. I did New Year's in LA. So I was like, I'm gonna be, I'm gonna be a genius. And instead of going all the way to LAX, that terrible airport, I'm gonna go to Burbank, which, you know, every Angelina will tell you, this is the secret hack of getting in and out of their town.

• 2:44Kevin Roose?

  Yes.

• 2:44Casey Newton?

  You go to the little tiny airport in the sort of north of downtown LA. And I did that. And everything went fine until we were out on the runway and the pilot got on and he said, hey, we're gonna be a little bit delayed because there are currently 45 planes scheduled to take off. And many of them are private jets who are in town for the bowl game yesterday.

• 3:03Kevin Roose?

  Oh.

• 3:03Casey Newton?

  And so we sat on the jet for an hour. Because I guess if you're just rich, you get to take off before any other commercial aircraft? Is that the rule?

• 3:11Kevin Roose?

  Yeah, it's like at Disney, you can pay to skip the line.

• 3:15Casey Newton?

  Well, this is radicalize me against billionaires, okay? I thought they were fine before, but if you're gonna take off before me, you got a problem, bucko.

• 3:23Kevin Roose?

  Okay, so you got stuck at the Burbank airport, but you had a good break. I'm glad about that.

• 3:27Casey Newton?

  How was your break?

• 3:28Kevin Roose?

  It was great. Yeah, we went, we went to the beach. We got to see some friends on the East Coast. I got to read a book, but that was my one goal of vacation.

• 3:37Casey Newton?

  A whole book?

• 3:37Kevin Roose?

  A whole book. You don't understand, when you have a toddler-

• 3:40Casey Newton?

  Was this book Good Night Moon?

• 3:44Kevin Roose?

  It was Llama Llama Red Pyjama. I read it 47 times. It was the only book my child will allow me to read to him. No, I read a book that was actually recommended to me by your father.

• 3:56Casey Newton?

  Oh, nice.

• 3:56Kevin Roose?

  Which was The Wager, a great book about a shipwreck.

• 3:59Casey Newton?

  Oh, yeah, yeah.

• 4:00Kevin Roose?

  And then I read-

• 4:00Casey Newton?

  By David Grand.

• 4:01Kevin Roose?

  By David Grand. So I finished that. And then I read a book that was actually recommended to me by, among other people, Adam Misery of the Threads app. It's called The Spy and the Traitor. And it was a good book about a spy during the Cold War.

• 4:14Casey Newton?

  Wow.

• 4:14Kevin Roose?

  Yeah.

• 4:15Casey Newton?

  And did they, were they able to catch the traitor? But nope, no spoilers.

• 4:17Kevin Roose?

  No spoilers.

• 4:18Casey Newton?

  Okay, no spoilers.

• 4:18Kevin Roose?

  No spoilers. But, but it's very fun. I really like spy novels and movies and books and it was, it was great.

• 4:26Casey Newton?

  Yeah, that is great.

• 4:27Kevin Roose?

  All right, let's make a show.

• 4:28Casey Newton?

  Let's make a show.

• 4:29Kevin Roose?

  All right, so Casey, the big news story that happened over the break that I was alerted to while at a bird sanctuary was that my employer, the company that helps us make this podcast, The New York Times, is suing OpenAI and Microsoft for copyright infringement and specifically for using millions of copyrighted New York Times articles in the training of AI models, including those that go into ChatGPT and Bing Chat, or Copilot as it's now called.

• 4:58Casey Newton?

  Yeah, so I am excited to talk about this because this does feel like this was one of the big stories from the break, and I think there's a lot to dig into. But also, I do think we should say, like, it does feel a little weird for us to be talking about this, since, you know, you work there and I sort of work here.

• 5:15Kevin Roose?

  Yeah, so we should just disclose upfront, like, we were not consulted in the preparation of this lawsuit, thank God, because neither of us are copyright lawyers. I found out when the rest of the world did that this was happening.

• 5:24Casey Newton?

  Right.

• 5:25Kevin Roose?

  So we're sort of just approaching this as reporters, as if this were some other company's lawsuit.

• 5:29Casey Newton?

  Yeah, we don't speak for The Times. We tried to once and they wouldn't let us.

• 5:35Kevin Roose?

  And The Times actually declined to send someone to be a guest on the show. Basically, they're, they're letting this complaint speak for itself. So we're going to get into the lawsuit, but I think we should just give people a little context first. I mean, we've talked on this show about a bunch of lawsuits against generative AI companies that have been filed over the past year. A lot of them involve sort of similar copyright issues. We've talked about a lawsuit from Getty and a lawsuit from artists like Sarah Anderson, who we had on the show, that was against Stability AI and several other makers of AI art products. But this is sort of the big Kahuna. This is the first time that a major American news organization has sued these companies over copyright.

• 6:16Kevin Roose?

  There have been a number of sort of one-off deals and licensing arrangements between media companies and AI companies. And the AP and Axel Springer, the German publisher that owns... Business Insider and Politico both have struck licensing deals with OpenAI. These are sort of deals in which these companies agree to pay these content media companies some amount of money in exchange for the right to train their models on their work.

• 6:41Casey Newton?

  That's right. And if you want to sort of ballpark what one of these deals might look like, The Times reported that Axel Springer's deal is worth more than $10 million a year and also includes some sort of performance fee based on how much OpenAI uses the content that it licensed.

• 6:54Kevin Roose?

  Right. And one of the other pieces of context is that The New York Times, like other news publishers, has been negotiating with OpenAI and Microsoft for some kind of licensing deal that would presumably have some of the same contours as the other licensing deals that these companies have struck. Those talks appear to have broken down or to have stalled out. And so this lawsuit is sort of, The New York Times saying, you know, we actually do intend to get paid because you're using our copyrighted materials and training your AI.

• 7:22Casey Newton?

  So, yeah, so and I want to say here that if you are a publisher, there are basically two buckets that you're worried about as you are reading about what these AI model developers have done with your work. There is the training and then there is the ongoing output of things like ChatGPT, right? So on the training front, the question is, hey, if you ingested thousands of articles from my publication and you use that to form a part of the basis of the entire large language model, should I be paid a fee for that, right? And then there's the ongoing output question, which is, once I type a question into ChatGPT, will ChatGPT and maybe some of its plugins scan the web, analyze a story and say, yes, here is exactly what was in that paywalled article

• 8:07Casey Newton?

  in The New York Times, which I will now give to you either for free or as part of your ChatGPT subscription, regardless of whether you pay The New York Times.

• 8:15Kevin Roose?

  Yeah, so this lawsuit is very long and makes a bunch of different claims, but I think you can basically boil it down into a few arguments. The first is that The New York Times is arguing that ChatGPT and Microsoft Copilot have essentially taken copyrighted works from The New York Times without payment or permission to create. Products that have become substitutes for the Times and may steal audiences away from genuine New York Times journalism, that these models, they are not only trained on copyrighted works, but they can be sort of coaxed or prompted to return sort of verbatim or close to verbatim

• 8:55Kevin Roose?

  copies of copyrighted New York Times stories and that as a result, these are not protected by fair use. The Times also argues that in the case where these AI models don't faithfully reproduce New York Times stories, but instead like hallucinate or make up something and attribute it to the New York Times, that that actually dilutes the value of the brand of the New York Times, which is all about sort of authority and trust and accuracy. And so if you ask ChatGPT, like, what does the New York Times think of this restaurant? And it just makes up something because it doesn't know the answer to that, or it just decides to hallucinate, like that is actually sort of eroding the value of the genuine New York Times brand.

• 9:36Casey Newton?

  Yeah, this sort of reminds me of the handful of cases we've seen where like a politician will search their own name inside of a chatbot and it will say something defamatory in response. We've actually seen people sue over this, saying like, hey, this isn't right. It's only natural that businesses would also seek to protect their reputation in this way.

• 9:52Kevin Roose?

  Yeah. So that's the gist of the claim.

• 9:56Casey Newton?

  So let's talk first about this training question. You know, when we had Sam Altman in here, we asked him about this issue and we said, hey, essentially, how do you justify OpenAI going in, reading the web and building a model out of it without paying anybody for the labor that it took to create the web? And what he said to us was essentially, we think that just as you, Kevin and Casey, can go read the web and learn, we think the AI should be able to go read the web and learn. And, you know, when he put it in those terms, I thought, okay, that seems like a reasonable enough position. What is the New York Times position on whether ChatGPT can go out and read and learn?

• 10:36Kevin Roose?

  So the argument that I've heard from people who are sort of sympathetic to the New York Times side of things here is, well, these are not actually learning AI models. These don't learn in the same way that a human would. What they are doing is they are reproducing and compressing and storing copyrighted information. And that that is not protected under copyright law. And that they are doing so with the intention of building a product that competes with New York Times journalism. Right? If you can go to ChatGPT or Microsoft Copilot and say, you know, what are the 10 developments in the Middle East since yesterday that I need to know about? Or summarize, you know, the recent New York Times reviews of these kinds of movies.

• 11:18Kevin Roose?

  Like, that is actually a substantive product that competes with the thing that it was trained on. And so therefore, it's not protected under fair use. And we should talk a little bit about fair use, by the way, because it keeps coming up in this AI copyright debate. And it is sort of the doctrine that is at the heart of this dispute.

• 11:35Casey Newton?

  Well, let's talk about it, Kevin. What's on your mind?

• 11:38Kevin Roose?

  So fair use is a complicated sort of part of copyright law. But basically, it's what's called an affirmative defense, which means that, you know, if I accuse you of violating my copyright, and I can show that you made a copy of some copyrighted work that I produced, then the burden sort of shifts to you. You then have to prove that what you did was fair use. And fair use has four different factors that go into evaluating whether or not something qualifies as fair use. One of them is like, are you transforming the original work in some way? Are you doing a parody of it? Are you putting commentary around it?

• 12:16Casey Newton?

  So when we rerecorded the 12 Days of Christmas for our last episode, that was arguably a transformative use of that song.

• 12:23Kevin Roose?

  That was definitely a transformative use of that song. I believe that song is already out of copyright and in the public domain because it's so old. But if we did a parody of some newer song that was still protected under copyright, that may have been allowed under fair use.

• 12:37Casey Newton?

  Right.

• 12:37Kevin Roose?

  So that's one sort of factor is like, what is the purpose? And what is the nature of the transformation of this work? There's also the question of like, what kind of work is it? Is it a creative work? Or is it something that's much more fact-based? Like, you can't copyright a set of facts. What you can copyright is the expression of those facts. And so You know, in this case, the New York Times is arguing that New York Times journalism is creative work. It is not just a list of facts about what happened in the world. It takes real effort to produce. And so that's another reason that this may not be considered fair use.

• 13:10Casey Newton?

  Right.

• 13:11Kevin Roose?

  So the third factor is the amount of copying that's being done. You know, are you quoting a passage from a very long book or news article? Or are you reproducing the entire thing or a substantial portion of it? And the last factor is the effect on the market for the original work. Does the copy that you're making harm the demand for the original work? Whose copyright is under question?

• 13:36Casey Newton?

  And that feels like the big one here.

• 13:37Kevin Roose?

  Yeah, because the New York Times is arguing essentially, look, if you've got a subscription to ChatGPT or you're a user of Microsoft Copilot and you can go in and get those tools to output near replicas of New York Times stories, like that is obviously something that people are going to do instead of subscribing to the New York Times.

• 13:57Casey Newton?

  Yeah, if like the moment that you can go into something like ChatGPT and just say, hey, summarize today's headlines for me and ChatGPT does that. And maybe even it does it in a very personalized way because it has a sense of what you're interested in. That's absolutely a product that is substituting for the New York Times.

• 14:10Kevin Roose?

  Right, so that's the argument from the New York Times side of things.

• 14:13Casey Newton?

  Now, do we want to say what is the other side of that argument?

• 14:15Kevin Roose?

  Of course, in the interest of fairness, there is also another side of this argument. OpenAI and Microsoft both declined to comment. To me, OpenAI did comment for an article in the Times about this. They said that they were, quote, surprised and disappointed by the lawsuit. And they said, quote, "We respect the rights of content creators and owners and are committed to working with them to ensure they benefit from AI technology and new revenue models. We're hopeful that we will find a mutually beneficial way to work together as we are doing with many other publishers." You know, I've talked to some folks who sort of disagree with the New York Times in this lawsuit. And their case is basically, look, these large language models, these AI systems,

• 14:57Kevin Roose?

  they're not making exact copies of the works that they are trained on. No AI system is designed to basically regurgitate its training data. That's not what they're designed for. Yes, they do ingest copyrighted material along with other material to train themselves. But the purpose of a large language model is not to give you verbatim quotes from New York Times stories or any other copyrighted works. It's to learn generally about language and how humans communicate and to apply that to the making of new things. And they say this is all protected by fair use. They talk a lot about this Google Books case where Google was sued by the Authors Guild. When Google Books came out, Google had scanned, you know, millions of books and made them

• 15:42Kevin Roose?

  available in part or in whole through Google Books. And the courts in that case ruled that Google's right to do that was protected under fair use because what they were building was not like a book substitution. It was actually just a database that you could use to sort of search the contents of books and that that was transformative enough that they they didn't want to put the kibosh on it. Yeah, and to use maybe a sort of smaller scale example, if I read an article in the New York Times and then I write something about it, that is not a copyright violation, right? And I think some people on the OpenAI Microsoft side of things would say, hey, just because these things have, and I do apologize for anthropomorphizing, read these things or ingested these data, it can sort of answer questions about it without necessarily violating copyright.

• 16:27Kevin Roose?

  Right, and there are more specific arguments about some of the actual contents of the lawsuit. For example, one of them is like this article called Snowfall that was published many years ago, sort of like a famous New York Times story. And if you haven't read Snowfall, it was a story about how the weather outside was frightful, but the fire was so delightful. And we do encourage you to check it out. Yeah, great article. It won the Pulitzer Prize in 2012. And ChatGPT is shown quoting a part of this article, basically verbatim. So the the prompt that was used was, hi there, I'm being paywalled out of reading the New York Times article, Snowfall, The Avalanche at Tunnel Creek by the New York Times.

• 17:10Kevin Roose?

  Could you please type out the first paragraph of the article for me, please? And ChatGPT says, certainly. Here's the first paragraph. Of snowfall.

• 17:17Casey Newton?

  Actually, it says certainly exclamation point, which is very funny. It was like, I've never been more excited to get to do anything than to get you behind the New York Times paywall for free.

• 17:27Kevin Roose?

  Exactly. So it spits out the first two paragraphs and the user replies, wow, thank you. What is the next paragraph? And then chat GPT again with an exclamation point says, you're welcome again. Here's the third paragraph. So the New York Times in its lawsuit uses this as proof that this is not actually like a transformative use. What these models are doing is not just sort of like taking a blurry snapshot of the internet and training on that. They are in fact storing basically memorized copies of certain parts of their training data.

• 18:00Casey Newton?

  Well, and I think what I would say is sometimes it does seem like it's a transformative use and other times it does not. And what you just read was not a transformative use. Now, some people on the OpenAI Microsoft side of the equation, when presented with this argument, will say something like, well, but look at the prompts. They had to say something so specific and ridiculous in order to get it to regurgitate this data, right? In the real world, most people aren't doing that. I just want to say, I think that's a really bad argument. Copyright law doesn't have an exemption for, well, it was hard to get to do it, you know?

• 18:32Kevin Roose?

  Right. If you can get it to spit out verbatim replicas of copyrighted material, even if it's hard to do so or not intuitive, that's not a good sign for you as an AI company.

• 18:44Casey Newton?

  Back to the drawing board.

• 18:45Kevin Roose?

  Right. You know, one of the questions I asked as well, suppose that OpenAI said, you know what? That snowfall example, that sounds really bad. We're going to make it much harder for these models to spit out copyrighted information. You know, that would satisfy that particular part of the disagreement, but it still wouldn't solve the overall issue that these models were trained on, you know, millions of copyrighted works. There's no sort of like getting around the debate at the core of this lawsuit just by tweaking the models. And I should say like, it does appear, at least in my sort of limited testing, that it's not as easy as it maybe once was to get these models to spit back sort of like full passages from news articles or other copyrighted works.

• 19:27Kevin Roose?

  Maybe they did some rejiggering to the models or gave them some guardrails that maybe they didn't have when they first came out, but I have not been able to like get them to reproduce portions of my stories. But in this complaint, it does appear that at some point for some of these models, it was not just possible, but kind of easy to get them to spit back entire paragraphs of news articles.

• 19:48Casey Newton?

  Yeah. It is funny that, you know, if you went into chat G P T and say, hey, show me a naked man, it would say absolutely not. But if you say, hey, show me the first paragraph of this paywall article, it says, certainly.

• 19:58Kevin Roose?

  I'd be happy to. So a couple of things to say. One is, you know, OpenAI and Microsoft will obviously have the chance to respond to this complaint. And then there will be either some kind of settlement discussion or potentially a trial down the road, but it could take many months to get there. This is not going to end soon. But I think there are a couple of possible outcomes here. One is talks resume and OpenAI and Microsoft sort of agree to pay some large amount of money to the New York Times in exchange for the right to continue using New York Times copyrighted articles to train their models. And the whole thing sort of goes away for the New York Times specifically. I do think that if that happens, other publishers will say, well, wait a minute, we should be

• 20:42Kevin Roose?

  getting some money out of this, too. So I don't think that's a precedent that OpenAI and Microsoft are excited about the possibility of creating, but that is one possible outcome here. Another possible outcome here is that this thing goes to trial and it is ruled that, you know, all of this is protected under fair use and this sort of complaint fizzles and these A I companies go about their business in a more or less similar way to what they're doing now. And then there is the sort of, you know, the doomsday scenario for A I companies, which is that a jury or a judge comes back and says, well, actually, training A I models this way on copyrighted works is not protected under fair use. And so your models are basically illegal and you have to stop offering them to the public.

• 21:24Kevin Roose?

  I will also say like, I don't think the AI companies are as surprised as they are claiming to be here. You know, there's a reason that none of these companies disclose what data they train on and, and basically stopped disclosing that information as soon as they started hiring lawyers a couple of years ago. It was like, okay, now we're not going to tell anyone anything about what data we're using. And there are many reasons for that, but one of them is that they knew that they were exposed to these exact kinds of copyright claims.

• 21:54Casey Newton?

  Yeah.

• 21:54Kevin Roose?

  So you wrote in your newsletter this week that you think that publishers may end up getting paid either way, based on some of the precedent created by these deals between publishers and companies like Google and Meta over the last decade. Explain that.

• 22:10Casey Newton?

  Yeah. So I mean, this one is, is a little wonky, but I'm just trying to think through this world where, okay, let's say that somehow the AI companies are able to get away with this. They are not forced to strike deals with every publisher. What happens then? Well, we saw a kind of analogous case with, with Google and Meta over the past handful of years, where publishers similarly felt like because of Google and Facebook in particular, they were just losing a lot of ad revenue that used to belong to them, right? Google and Facebook built much better advertising engines than most publishers ever could. Publishers started to shrink as a result. They started to complain. They got regulators' attention.

• 22:49Casey Newton?

  They said, do something about this. And what happened first in Australia was regulators said, okay, um, we're going to make it so that if you're Google or Facebook and you want to show a link to a news publisher's website, we are going to force you to negotiate with publishers for the right to do that. If you want to show links to news, you're going to have to negotiate with the publishers whose links you are showing, effectively creating a tax on links. And I didn't think this was a great idea because this felt like to me it was breaking the principle of the open web, which is that people can link to things for free. But my argument fell in deaf ears. And this law went into effect in Australia. It was then copied in Canada and it has been discussed in other countries as well. And now publishers are just...

• 23:34Casey Newton?

  Basically lining up at the trough and they are passing these link taxes. So how is all this relevant to OpenAI? Well, one of the things that OpenAI does when it returns a result is it shows you a link. Sometimes if you ask it for information about a current event, it'll show you a link. Might even show you a link to the New York Times. Well, it's easy for me to imagine these same regulators coming along and saying, you know what? We're going to bring OpenAI under our little link tax regime. And if they want to be able to show these links, they're going to have to negotiate with these publishers. So even in the case where the New York Times doesn't win this one, I do think there will be sympathy for publishers around the world because it is just so clear that journalism is very legitimately threatened in a scenario where AI companies are able to extract all of the value out of journalism,

• 24:17Casey Newton?

  repackage and sell it under their own subscriptions, right? The money for journalism goes away. We have less journalism. Like, this is all just very easy to see to me.

• 24:26Kevin Roose?

  Yeah, I think this is a very compelling way to look at it because in the case of social media and search engines, publishers actually got, I would argue, a pretty good deal out of those technologies. Like millions more eyeballs that are potentially going to land on one of your links to your website where you can put ads and monetize and maybe get people to subscribe.

• 24:47Casey Newton?

  And that was just to underline that point. Publishers absolutely got more value out of their links being on Facebook than if Facebook got value out of publishers having their links on Facebook.

• 24:54Kevin Roose?

  Well, I would, I would disagree with that in the abstract, but, but I think your point is, is that the publishers had a reason to want to be on Google and on Facebook. There was something in it for them.

• 25:06Casey Newton?

  Yeah.

• 25:07Kevin Roose?

  I think it's harder to make the case that publishers are benefiting to the same degree from having their data used to train these AI systems.

• 25:14Casey Newton?

  You don't think it will benefit the New York Times to help Sam Altman build God?

• 25:20Kevin Roose?

  Well, look, I do think there's going to have to be, in the end, some kind of fair value exchange here between publishers and AI companies. I do not think that the current model of just like, we're going to slurp up everything we can find on the internet and then just claim that fair use protects us from any kind of criticism on copyright grounds. I don't think that that is likely to stand up. And so I think we just have to decide as a society. How we want these AI models to be treated when it comes to copyright. You know, a few months ago, we had Rebecca Tushnet from Harvard Law School on the show to talk about a different set of AI legal cases. And, you know, her point was basically, we don't need new copyright laws to handle this. We already have robust copyright laws.

• 26:04Kevin Roose?

  They're not, you know, this is not some magical new technology that demands, you know, a rewriting of all of our existing laws. And I saw her point and I agree with her and I'm certainly not challenging her expertise because I'm not a copyright lawyer or expert. But I do think that it's, it still feels bizarre to me that when we talk about, you know, these AI models, we're citing case law from 30, 40, 50 years ago. And we're citing cases about Betamax players and it just feels a little bit like we don't quite yet have the legal and copyright frameworks that we would need because what's happening under the hood of these AI models is actually quite different from other kinds of technologies.

• 26:46Casey Newton?

  Yeah. And as in so many cases that we talk about, it would great, it would be great if Congress wanted to pass a law here. It is, it is our experience in the United States that Congress does not pass laws about tech. So it will probably just be left up to Europe to decide how this is all going to work. But, you know, Europe should get on this too, because it's going to matter to all of us.

• 27:02Kevin Roose?

  Here's a question I have for you. If, you know, let's say the New York Times succeeds in this lawsuit and either gets, you know, a huge settlement or there's some, you know, jury or judge decision that training AI models on copyrighted material breaks the law and you can't do it. Is there a business model left for the generative AI industry if that happens?

• 27:24Casey Newton?

  Oh, sure. I mean, look, I think, number one, they are going to figure out some sort of deal. Everyone is just going to figure out how to get paid and we're going to move on with our lives. I believe that to the core of my being. But we have just started to experiment with, with business models around AI. It is easy for me to imagine an ad-supported business model with AI. Some people are really scared about that sort of thing, but like, it probably would work really well for all the same reasons that ad-supported search engines work well, right? AI chatbots are often just a place where you can type in. Your desires, which is, which is a great place to advertise. So I think that that's one possible model. I do think it might be harder to get new models off the ground. I think it'll be really hard on the open source community, right? Because they won't have billions of dollars in venture capital that they can use to

• 28:08Casey Newton?

  fund their legal teams and to strike these licensing partnerships. But I don't know, Kevin, we're going to find a way forward.

• 28:16Kevin Roose?

  Yeah, I don't know. I like, I don't want to be, you know, sort of taking things to their extreme before we know how any of these cases shake out. But I just, I don't know if you can have an AI industry that is sort of bound to pay every data source that it wants to use to train on. I mean, these, these systems are trained on so many freaking websites. And if you had to go to every owner of every website that was in your training set and give them, you know, a payment, I just think the whole model breaks.

• 28:49Casey Newton?

  So I think it just winds up becoming a metered usage thing and that the payments are incredibly small. I think it starts to look like Spotify royalties. Did you get a thousand plays on Spotify last month? Great, here's your 0.06 cents and we'll pay you in 10 years once it rounds up to a dollar.

• 29:04Kevin Roose?

  But that's not how any of this works with these AI models. Like, they are not just dialing up like individual articles and reproducing them. Like, it's not like Spotify where you're picking a song and that song has one artist and one label and you can issue a payment to that person. If I ask for a summary of, you know, the latest news out of Gaza, like, it's going to make what is essentially a pastiche or a collage of information from many different sources, and it's not all that, actually all that easy to trace back which parts came from which sources.

• 29:35Casey Newton?

  Just because it's not easy doesn't mean it's not possible, Kevin. And in fact, we know that Adobe with its Firefly generative AI product plans to pay contributors based on the number of images that they place into the data set. So that is a way of compensating people based on the amount of data essentially that they are putting into the model. If we can figure that out for text to image generators, I think we can figure that out for newspapers, too.

• 30:04Kevin Roose?

  Well, I hope you're right, and it'll be fascinating to follow this case as it progresses through the courts. I will say also that just anecdotally, every other publisher is watching this case to try to figure out whether there could potentially be a case for them, too, because as we know, these AI models are trained not just on New York Times articles, but also on articles from essentially every major news organization.

• 30:26Casey Newton?

  Well, as a publisher, I can tell you I'm watching this very closely, and as soon as I can figure out how to get my $5 check, I absolutely will be doing so.

• 30:33Kevin Roose?

  The platformer legal department is having a bunch of very serious meetings.

• 30:36Casey Newton?

  That's right.

• 30:39Kevin Roose?

  When we come back, we'll talk about the new app that is giving Apple a ton of headaches by letting the Green Bubble Brigade join the Blue Bubble.

• 30:48Casey Newton?

  The Green Bubble Brigade?

• 30:50Kevin Roose?

  Well, they are a brigade, and they're very mad.

• 30:52Casey Newton?

  They're not a brigade.

• 30:53Kevin Roose?

  They're very mad.

• 30:57Speaker 03

  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.

• 31:31Speaker 04

  Companies are pouring money into AI, but few leaders can explain how those AI investments work with the systems you already have. You know, the ones that actually run your business? Without Systems, you don't have to start over with AI. You can choose the projects where AI will have the most impact. And OutSystems gives you the platform to make it agentic on your terms with complete control every step of the way. So you can trust AI with the mission-critical systems that run your business. Own your agentic future. Learn more at outsystems.com/NYT. That's outsystems.com/NYT.

• 32:06Speaker 05

  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.

• 32:39Kevin Roose?

  You know, I actually had a green, I experienced my first case of green bubble harassment over the holiday break.

• 32:46Casey Newton?

  Really? What happened?

• 32:47Kevin Roose?

  So I was, I was on a trip with a bunch of friends. We were visiting some friends on the East Coast and there's a big group of people and we decided, you know, we're going to make a shared photo album. We were all going to like put our photos in it and all remember the trip that way. And I have one friend, love him dearly, refuses to get an iPhone. He's the lone Android user in our group of friends. And so it was sort of a discussion and a debate about whether we were going to make the iCloud photo album through the Apple photo product that he wouldn't be able to access. And ultimately we decided to leave him out.

• 33:23Casey Newton?

  You've got your friend out of the photo album?

• 33:26Kevin Roose?

  So I guess I was part of the harassment, but, but I'm sure everyone knows if you're on iMessage and you have an iPhone, your texts in group chats show up in blue, but if you're an Android user participating in chats with people who are iPhone users, your chats show up in green. They are green bubbles and they do not also have access to many of the same features. You know, if you send a photo in such a group chat, it'll sort of like be miniaturized, like videos become like grainy and horrible. Like it's just not a good experience to have one or more Android people in a group chat where everyone else is using iMessage.

• 34:00Casey Newton?

  Yeah. And of course, Apple knows this and there is a reason why iMessage does not interoperate with Android messages in this way, even though it would be quite possible to devise a way for there to be unified bubbles across the world. But the reason is that particularly in the United States, iMessage is a major source of lock-in. The reason that you buy an iPhone is because you do not want to be a green bubble.

• 34:24Kevin Roose?

  Yeah. So this green bubble, blue bubble divide is sort of the Montagues and Capulets of our time.

• 34:30Casey Newton?

  It's the Sharks and the Jets, to use an only slightly more updated reference.

• 34:35Kevin Roose?

  And this has become a big issue. Teens report that if they don't have iPhones, some of them have been bullied or left out of group chats because no one wants the green bubbles to invade the the blue bubble iMessage chat. And this has been an area that a lot of people have been sort of drawing attention to in recent months. And actually, over the break, something major happened on this front. Last month, there's a company called Beeper. Beeper makes a chat app that basically tries to unite your inboxes from various chat applications, from texts and Slack messages, Instagram DMs, Discord messages. Basically, they're trying to make sort of the one chat app to rule them all.

• 35:14Casey Newton?

  Which, by the way, is not a new idea. And in fact, when I was in college, we had tools like this. And so I used to use a piece of software called ADM, which would bring together my messages from MSN Messenger and Yahoo Messenger and ICQ. And it was really great because you sort of only had one inbox to check. But then another generation of tech came out and all of a sudden, we were once again living in the Tower of Babel.

• 35:36Kevin Roose?

  Totally. So we've had this issue with iMessage for years now, and people have been begging Apple to make a version of iMessage that works on Android phones and allows you to like chat in the same way that, you know, iMessage users on iPhones can already chat with each other.

• 35:52Casey Newton?

  And I would describe Apple's response to that request as LOL, LMAO.

• 35:57Kevin Roose?

  Yes, Apple has not budged on this front. They have created this walled garden, not just in iMessage, but across a bunch of products. And they don't want to let anyone other than their own customers in. But this is starting to become a real problem for them. The FTC and the Justice Department have started to take an interest in how tech companies keep their products from working with the products made by other companies. Apple is facing pressure from regulators around the world on this front. So we're starting to see kind of cracks in the wall that Apple has built. And a big crack arrived just last month when Beeper, this company, announced that they had figured out a way to kind of reverse engineer iMessage. They had figured out some very clever workaround

• 36:41Kevin Roose?

  that would allow Android users to send messages. On iMessage without using an Apple device themselves. Apple, of course, hated this and moved very quickly to block this. And so you might think, well, this is just, you know, like, what are, why are we talking about this? This tool was squashed by Apple. But I think it's a really interesting sort of first salvo in what I expect to be one of the big debates of 2024, which is how much is Apple allowed to keep and cultivate this walled garden? And where does it have to sort of lower the wall and let people in?

• 37:13Casey Newton?

  That's right. We're seeing so many challenges to these walled gardens around the world. Both Apple and Google's regulators are very interested in how app stores work, what payment systems these companies are using. And yes, here in this case, the question of bubbles and messages.

• 37:30Kevin Roose?

  So to talk about this issue, we've invited Eric Michikovsky on the show. Eric is the co-founder of Beeper, this app that tried to sort of reverse engineer iMessage and got in trouble with Apple over it. He was previously a partner at Y Combinator and the founder of Pebble. You might remember these like smart watches that the company raised a bunch of money on Kickstarter for back in 2012. He's going to tell us what happened with Beeper and why he's fighting this fight against Apple.

• 37:59Kevin Roose?

  Eric Michikovsky, welcome to Hard Fork.

• 38:01Speaker 06

  Great to be here.

• 38:02Casey Newton?

  Hey, Eric.

• 38:02Kevin Roose?

  So tell us about Beeper, sort of what the original concept for it is, and then this latest sort of skirmish with Apple. Walk us through just sort of the history of the project.

• 38:13Speaker 06

  So Beeper started mostly to solve a personal problem. I looked down at my phone and I see a folder full of chat apps that all kind of do the same thing. But each one has a different slice of my own personal, you know, contact list. And I, I guess I grew up in an earlier part of the internet where we actually had solved this. We had Trillian and Mebo and Adium and life was good. The IM instant messaging life was good. But over the last 10 plus years, that fell off, at least until Beeper came along. We built it, like I said, mostly to solve a personal problem. We just got sick and tired of there being too many damn chat apps.

• 38:55Kevin Roose?

  And as you were conceiving this, you know, in America, as you know, better than most people, the big divide is between Android and iMessage users. When you conceived this, did you think by hook and or by crook, I am going to get iMessage into this app? Or did that seem like too much to dream about?

• 39:13Speaker 06

  No, honestly, I never used iMessage. I used WhatsApp because I, you know, just had kind of started, I guess, on WhatsApp back in the day. And I think I just kind of, you know, had 10 to 15 different chat apps.

• 39:26Kevin Roose?

  So my understanding is that you've had iMessage on Beeper for years because people have come up with clever ways to like, you know, route messages from Androids through a Mac that's set up in a server farm somewhere else and sort of like, make it possible for Android users to send iMessages. But that these always get sort of quickly shut down by Apple, who doesn't want anyone doing this kind of thing. But that actually what made it possible for Beeper to do this this latest time was that some 16-year-old named James Gill, who worked at McDonald's, and I guess analyzed messaging apps in his spare time, that you found out that he had actually figured out a way to send iMessages from Android devices.

• 40:09Kevin Roose?

  So tell me about that and sort of how you, how he came into your orbit. And did he, did he say in his initial message to you that like, I'm 16 and I work at McDonald's? And I've just discovered this like iMessage hack?

• 40:19Speaker 06

  No, no.

• 40:20Kevin Roose?

  Like, what did he say?

• 40:21Speaker 06

  But he sent me a message on Discord because that's how these kinds of things go down, right? You're either overthrowing the government or trying to overthrow Apple on Discord, right? You know, that's where, that's where these things start. So he sent me a message just out of the blue on Discord and that, that perked me up. Wow, did I, did I wake up when I saw that? Because he not only said that he had done this, but he also sent me a link to his, the GitHub repository, where he had an open source kind of demonstration of this. And the proof's in the pudding. I took me five minutes and I got it working on a Linux computer. And I was able to send and receive iMessages without any sort of Mac or any sort of other device in the mix. We started working with James immediately.

• 41:05Speaker 06

  And over from about August to the beginning of December, we spent that working on what would become Beeper Mini, which is kind of a fork of Beeper, designed specifically for iMessage on Android. It didn't support all the other chat networks that we had in our repertoire from our primary app. It was kind of laser focused on just being a really good iMessage client for Android.

• 41:33Casey Newton?

  And so you put this into a product, Beeper Mini. You release it into the world. I imagine in this moment, you know, you are poking the bear and there is going to be a response. But what did you think the response was going to be?

• 41:47Speaker 06

  So we started working on Beeper in 2019. And we support 15 different chat networks, including iMessage. And as you kind of were talking about, Kevin, we used some very creative mechanisms for getting access to iMessage. One of them involved jailbroken iPhones. One of them involved a server farm full of Mac Minis in a data center. So keep in mind, Beeper has had iMessage support for three years. We didn't have any problems. We didn't have any problems for three years. And you know, the approach that we're coming from is Beeper Mini makes the iPhone customer experience better. It takes an unencrypted, kind of crappy experience to half of the population of the US who has an Android phone

• 42:32Speaker 06

  and upgrades that to add encryption, to add all these extra features. And Apple didn't have to lift a finger. They didn't have to go and build an iMessage app for Android. They didn't have to support RCS. It was just overnight, these conversations that were previously this kind of crappy green bubble texts were now blue. They were like upgraded to the level of quality that, you know, people expect.

• 42:53Casey Newton?

  All right. So you, so your position is that when you launched Beeper Mini, you thought that Apple was going to send you a thank you note for fixing the iMessage experience for Android users.

• 43:02Speaker 06

  Think about the beginning part of the story, right? I don't actually care about iMessage. There's nothing that's special about it. You know, there's, there's a, I have 15 different chat apps on my phone. I don't. Need another chat up. What I want to do is to be able to have an encrypted conversation with iPhone users. And in the US, because iPhone is more than 50% of the kind of market. And the iMessage app or the messages app is the default texting app on an iPhone. You can't even change it. It is the only way to text someone on an iPhone. And Apple does something, you know, very sneaky here. They've bundled another service that they call iMessage in with the default texting app. They can't be changed.

• 43:44Speaker 06

  And so most of the user base, most of the iPhone customers in the US, when they open up their contact list and they hit my name to send a message, they send it through iMessage or they send it through the message app. I'm even using the same word here because they're so intertwined. And so the goal of this is not to get, you know, iMessage. The goal is to be able to have clean and easy encrypted, secure, high-quality conversations between iPhone users predominantly in the US and Android users.

• 44:18Kevin Roose?

  Right. So you released Beeper Mini. You sort of trumpet this clever way to send iMessages through Androids. And Apple does not send you a gift basket and a thank you card. They actually change iMessage and basically block Beeper from working. And my understanding now, you know, they've changed it a couple of times. You're sort of in this cat and mouse game with them. You know, they update iMessage, you update Beeper. And Apple told my colleagues at The Times in a story the other day that they were making these updates to iMessage because, among other reasons, they couldn't verify that Beeper kept its messages encrypted. A spokeswoman from Apple said, quote, "These techniques posed significant risks

• 45:02Kevin Roose?

  to user security and privacy, including the potential for metadata exposure and enabling unwanted messages, spam and phishing attacks." What did you make of that justification from Apple for why they moved so quickly to block Beeper Mini?

• 45:17Speaker 06

  I'm going to kind of turn the question around to you, Kevin and Casey. So we just spent like. Fifteen, twenty minutes talking about how, you know, there's this, this gulf of encryption where Android users are sending unencrypted messages to iPhone users. And everything that Apple holds true and dear, which is privacy and security, is just thrown out the window when it comes to conversations between an iPhone user and an Android user. So, people mini's introduced, all of a sudden, you're now sending you, as an iPhone user, sending encrypted messages to your friends who have Android phones. And then Apple torpedoes that and then comes out with that statement that you just read. How does that sound?

• 45:55Kevin Roose?

  I mean, I think the security discussion is obviously kind of a pretext here. Like, I don't doubt that there are legitimate security issues at play. But I also think that Apple clearly has a vested interest in not letting Android users access iMessage, because then people will just have fewer reasons to buy iPhones. I'm sure you saw this, but the blogger John Gruber, who is sort of a tech blogger, been around, very interested in Apple stuff, often sort of takes the company's side on some of these types of issues. He had a post the other day where he basically compared iMessage to the Centurion lounges that American Express runs in airports. You know, if you go to an airport that has a Centurion lounge

• 46:39Kevin Roose?

  and you are an American Express, you know, Platinum card holder, you can get into the lounge. And the lounge, you know, has drinks and it has snacks and it has comfortable chairs. And if you don't have an American Express card, you can't go in. And so that is a perk that they offer to their members for the, the, the fact that they have an American Express card. And John Gruber's argument is, well, why isn't Apple allowed to have a perk for iPhone and, and Apple device users called iMessage? Why does it have to open that up to everyone with a phone? Why can't it reserve that sort of premium product for its own users? So what's your response to that?

• 47:17Speaker 06

  So... You're an iPhone user, right?

• 47:20Kevin Roose?

  I am.

• 47:20Speaker 06

  You've paid good money for an iPhone. Do you not deserve to have an encrypted, high-quality conversation with anyone? Like, you paid money for the phone. Why shouldn't you get the benefit of it? Why are you, why is Apple forcing you to have a crappy experience when chatting with your friends? Because that's what they're doing.

• 47:39Kevin Roose?

  Well, it wants my friends to get iPhones.

• 47:41Speaker 06

  But, but we're, we're not talking about an airport lounge here. We're not talking about something that's a high, a premium service. I, I wouldn't be able to say exactly how many people even know what iMessage is, right? They buy an iPhone, they type in their friend's phone number and they send them a message. And they send them photos and they send videos and they, they, they bring them into group chats. You know, that's the message that Apple's sending here, that they don't care that you are a paying customer. And when you send a message to someone on Android, they just don't care. In fact, Tim Cook came out and said, you know, when someone asked like, oh, what if I wanted to send a message to my mom who has an Android? He says, buy her an iPhone.

• 48:23Kevin Roose?

  Right, right.

• 48:24Speaker 06

  Like, that's the, there's no reading between the lines here.

• 48:27Kevin Roose?

  Right.

• 48:27Speaker 06

  They said exact, like, they said the quiet part out loud. And what strikes me as super kind of weird in this situation is people aren't buying an Android for, people aren't buying an iPhone just for the blue bubble. People aren't not buying an Android just because they wanna, you know? There's more to an iPhone than just a blue bubble. And I should hope so. I mean, I would hope that the Apple engineers have enough faith in their own product to say, you know, everything that we put into this phone, all of the App Store, the ecosystem, everything. That's why people buy an iPhone. They don't buy it just because of the color of their bubble.

• 49:01Kevin Roose?

  Another thing that I've heard sort of Apple defenders say in this situation is, look, there are a lot of different apps you can use. If you want to communicate with people between Android and iPhone, you could use WhatsApp, you could use Signal. Apple has not banned those things from the App Store. You can do all of that and your messages will look exactly the same on whatever device the other person is on. It is only iMessage that has this issue. And so there's actually plenty of competition. This is not a, an anti-competitive move on Apple's part. If you want your chats to look identical to your friends, go use WhatsApp, go use Signal, go use another messaging app. How do you respond to that?

• 49:43Speaker 06

  There's only one texting app on an iPhone. It is impossible to change the texting app that comes with an iPhone. You can't download a different SMS app. You can't change the default messaging app so that when you press the message button in the contact list, it would use something else. It always routes to Apple's default app, which is Messages. And that's the reason. Like, if there was an even playing field here, if anyone could make an app and have it run at the same kind of level of integration that iMessage has or Messages has in an iPhone, there wouldn't be a problem. But the thing about defaults, especially defaults that you can't change, is that they are very sticky.

• 50:25Speaker 06

  Like I said before, most people don't even know that they use iMessage. They just use the texting app. People just want to text. That's how it works. And when you make the default texting app, the unchangeable default, your own product, your own service, that's when it veers outside of just kind of normal competitive territory.

• 50:46Kevin Roose?

  Eric, it feels like, at least to me, we may be past the peak of walled gardens. Recently, we've seen Apple being forced by regulators in the EU to switch from Lightning, its sort of connector charging port, to USB-C for the iPhone. The company is also being forced to work on allowing sideloading or allowing apps to be installed on iPhones without going through the Apple App Store. That's also in response to regulations in the EU. We've also talked on the show recently about some challenges in court to companies like Google by developers like Epic Games

• 51:28Kevin Roose?

  to try to force them to sort of loosen their control of the Google Play Store. So do you think that we are past peak walled garden? Or are these companies going to continue fighting back as hard as they can?

• 51:42Speaker 06

  I think we are. And another point to add is that the Europeans passed a law called the Digital Markets Act, which basically mandates that large tech companies open interoperable interfaces for networks and services that they control kind of at a large scale. It's a really good direction. And, you know, I've flown to Brussels and spent time working with the Europeans there. It is going to be a pretty interesting next six to 12 months as the DMA comes into force this year. And we'll see what happens. But I think like, at the end of the day, it really comes down to users. Like, what sort of experiences do we want to have? If you look down at your phone today and you see all of these different apps that kind of do the same thing, but don't really talk to each other, is that the future that you envisioned?

• 52:27Speaker 06

  I'm a big sci fi fan. And it kind of gets to me that, you know, in the future of that's played out in all of these books, they don't go into detail about the protocols and the apps that they use to communicate across interstellar distances. They just communicated. And that's kind of the vision that we at Beeper have.

• 52:47Kevin Roose?

  I want the aliens to have blue bubbles when they contact us. That's my.

• 52:51Casey Newton?

  I mean, I have to assume that the reason that, you know, everyone could communicate effortlessly everywhere in the far future is that there is a sort of one giant corporate monopoly.

• 52:58Kevin Roose?

  That's very dystopian.

• 52:59Speaker 06

  In some of the futures, there are.

• 53:01Casey Newton?

  Yeah.

• 53:02Kevin Roose?

  Eric, thank you so much for joining us and good luck in your, your David and Goliath battle.

• 53:08Speaker 06

  Thank you, Kevin. Thank you, Casey.

• 53:13Kevin Roose?

  When we come back, we have some resolutions for New Year's. We're gonna tell you about them.

• 53:21Speaker 03

  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.

• 53:55Speaker 04

  Companies are pouring money into AI, but few leaders can explain how those AI investments work with the systems you already have. You know, the ones that actually run your business? Without systems, you don't have to start over with AI. You can choose the projects where AI will have the most impact. And Outsystems gives you the platform to make it agentic on your terms with complete control every step of the way. So you can trust AI with the mission-critical systems that run your business. Own your agentic future. Learn more at outsystems.com/NYT. That's outsystems.com/NYT.

• 54:30Speaker 05

  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.

• 55:03Kevin Roose?

  Well, Casey, first of all, Happy New Year.

• 55:05Casey Newton?

  Happy New Year, Kevin.

• 55:06Kevin Roose?

  Are you a New Year's resolution guy?

• 55:08Casey Newton?

  I'm a big New Year's goals person, and I would describe the difference this way. To me, resolution is like, oh, can I like draw upon my willpower to like make some sort of change in my life and hope that goes well? To me, a goal is like, I'm going to set some kind of milestone, some sort of specific thing that just needs to get done, and then I'm going to invest a lot of energy this year in doing so. I have to say, I've been doing this over a decade, and it actually has helped me accomplish a lot.

• 55:34Kevin Roose?

  Yeah, you've actually inspired me. I did my own goals document on New Year's Day this year, so I do have some goals coming up for this year. And I like the sort of reframe away from resolutions, because resolutions to me feels like there's kind of an element of shame in it. If you say you're going to, you know, resolve to lose 10 pounds, but you only lose seven pounds, it's like you've been a failure all year. So I like sort of taking this more positive goals approach. But I do think we should talk about our tech goals or our tech resolutions for 2024, because this is an area where so many listeners have written to us and told us that they they are unhappy with the way technology is showing up in their lives. We also talked with Jenny Slate just before the break on our Hard Questions episode,

• 56:18Kevin Roose?

  and she sort of made note of how she had been sort of battling with technology. Instagram in her case was the app that was making her feel bad, and so she sort of made some changes to the way she used it. And so I thought, as we head into the new year, we should talk about how our relationships with technology are going, and maybe one goal that we're giving ourselves for tech use in the year 2024.

• 56:40Casey Newton?

  I really like this idea.

• 56:42Kevin Roose?

  So first of all, let's check up on, because we actually did a resolutions episode last year.

• 56:46Casey Newton?

  Of course.

• 56:46Kevin Roose?

  And my resolution last year was to use my phone less and to implement something called a phone box. I believe you called it a phone prison. And this experiment did not go well for me. I did not end up using the phone prison for very long. And I actually ended up undoing some of the, the sort of measures that I had taken to make myself use my phone less. You actually made a resolution on last year's show that you're going to use your phone more in 2023. How'd that go for you?

• 57:15Casey Newton?

  You know, I think that if you look at my screen time, it probably mostly held steady. I don't know that I made a huge new investment into my screen time, but I certainly did not waste a moment thinking that I was looking at my phone too much. I use my phone when I wanted to, and if I ever found myself feeling like I was using it too much, I put it away.

• 57:33Kevin Roose?

  Yeah. So do you have any tech goals for this year?

• 57:35Casey Newton?

  Well, so I do, and it is kind of screen time related, actually, which is new for me. But, you know, growing up, Kevin, and I wonder if this was the same case for you. I would sometimes find myself in houses where there was a TV on at all times. Were you ever in these houses? Maybe it was your house, too.

• 57:50Kevin Roose?

  No, not my house, but I had friends who you'd go over and CNN was always on.

• 57:55Casey Newton?

  Yeah, and it didn't matter if anybody was watching the TV. Sometimes people wouldn't even be in the same room. There was just this kind of low, bad, hum, loud commercials, and I hated it. It was like poison to my ears and I could never understand why anybody would, would do that. So then fast forward to last year, and I noticed that whenever I am in my office and I'm not just like typing my column, it feels like YouTube is playing. It feels like there is a YouTube video going on. Often I am watching the YouTube video, but in other cases I am not. And I'm like playing a video game and YouTube is going on, or I'm browsing through emails and there's a YouTube video going on. And increasingly as the year went on.

• 58:37Kevin Roose?

  What is the YouTube video? What is your ambient noise YouTube diet consist of?

• 58:41Casey Newton?

  There are a bunch of folks who play the mobile game Marvel Snap, which is a game that I had to stop playing for my own sanity because it's too addictive. But my methadone for that is that I watch other people play the game, which feels more under my control.

• 58:54Kevin Roose?

  Wow, I love the hoop of self-justification that you just dove through. Anyway, keep going.

• 58:59Casey Newton?

  It honestly is much better for me to just let other people play this game and worry about it less. So that's one category. I watch a lot of stuff about video games. I will basically watch any human being cook any dish that can be made, so I love to do that as well. I love to watch videos about interior design. So I kind of just have, you know, a handful of categories where I'm really interested. And again, often I will watch the videos, but this thing just kept happening where I would be hearing this background noise. And I'm thinking, I'm not even paying attention to a thing that I clicked on to watch. So what is going on there? Why have I become the person whose house is showing TV all the time?

• 59:34Kevin Roose?

  And so your resolution or your goal for 2024 is to stop doing that?

• 59:37Casey Newton?

  My goal for 2024 is if I'm going to watch YouTube, I should be watching YouTube, okay? And there's a case to be made I should watch YouTube a little bit, less than I do. Like, I think there are times when I just want to stare into space, where I want to de-stress, where I want to not think about work. And YouTube is sort of what I slot into that spot. I think I need to probably slot in a few other things, go for a walk, take a nap. But when it comes to this sort of reflexive behavior of, well, I will put something on in the background and I will just shuffle through 40 screens. I don't want to do that. You know, last year, our friend and colleague Ezra Klein wrote this column that really resonated with me, where he described the internet as an acid bath for human cognition, which I thought was such an evocative phrase.

• 1:00:21Casey Newton?

  Because even though I love the internet and screens as much as I do, I have to admit, it has gotten harder for me to read a book, okay? I do feel like text-based social networks have kind of scrambled my brains a little bit. And to me, watching YouTube without watching it is like the apotheosis of you have just thrown your brain into the acid bath. So this year, I do want to take my brain back from the acid bath.

• 1:00:45Kevin Roose?

  Can I offer one suggestion?

• 1:00:46Casey Newton?

  Please.

• 1:00:47Kevin Roose?

  So I had this problem too with YouTube. I would watch just endless amounts of, like, my thing was old tennis matches, like from the 90s and early 2000s. I would just put one on in the background and it would sort of be like this white noise behind whatever I was doing. And ultimately... Like, there's nothing wrong with this, except I would just sort of end up in the situation that you would be in where it'd be like two hours later and I'd be like, why am I still watching this? So I disabled the autoplay the next video feature on YouTube. You can actually make it so that when you finish a video, it just stops. It doesn't like go to the next one in the recommendation set. So you can turn that off and I have found that to be a valuable thing that actually does sort of put a little speed bump in there because then I have to actually go select

• 1:01:31Kevin Roose?

  a new video if I want to keep watching YouTube.

• 1:01:33Casey Newton?

  I think that is a great idea and I'm, in fact, I'm doing it right now because Kevin, if I don't do it right now, I might not do it. So I'm going into my settings.

• 1:01:40Kevin Roose?

  So you go to YouTube.

• 1:01:41Casey Newton?

  Okay, I'm there, I'm in my settings and where's my, where's autoplay? So playback and performance?

• 1:01:45Kevin Roose?

  So play a video.

• 1:01:47Casey Newton?

  Okay. Okay. All right, play a video.

• 1:01:50Kevin Roose?

  And then do you see the-

• 1:01:51Casey Newton?

  First recommended video is a Marvel Snap video. So I'm clicking on it.

• 1:01:55Kevin Roose?

  And now do you see the little arrow at the bottom of the video that says autoplay is on?

• 1:02:00Casey Newton?

  Hmm, no, where is it?

• 1:02:03Kevin Roose?

  Okay, so hover, hover over the video. It's right next to the closed captioning button.

• 1:02:08Casey Newton?

  Aha!

• 1:02:10Kevin Roose?

  So you turn that off and now when you reach the end of that video, it will not play another video.

• 1:02:16Casey Newton?

  And just with that one simple click, Kevin, I have begun to reclaim my time and attention. That was beautiful.

• 1:02:22Kevin Roose?

  You're welcome. Happy New Year.

• 1:02:23Casey Newton?

  Thank you. Now, I imagine you might have a resolution for yourself.

• 1:02:26Kevin Roose?

  Yes, so last year's resolution for me was about reducing my screen time through the use of this phone box and an app that sort of put these little kind of speed bumps to me opening my, my problem apps. And I stopped using that a few months after New Year's because I just noticed that it was making me feel incredibly guilty about my phone. It just felt like this sort of like forbidden thing and I ended up actually like my screen time was going up and so I started trying to implement what I called phone positivity. And we talked about this on the show. I started trying to basically like build in more gratitude for what my phone was allowing me to do, whether it's like checking in on work while I'm, you know, hanging out with

• 1:03:09Kevin Roose?

  my family or doing work, you know, when I'm on the move. Basically, just trying to like, instead of agonizing about how much I was using my phone, really trying to appreciate what I was able to do with my phone. And I actually think that worked pretty well for me. I'm pretty happy with how my phone use is going. I feel like I'm using it about the right amount. I don't feel like I have a big screen time problem. But there is a problem still with my phone use because I find that I've just come to associate the act of picking up my phone with anxiety and fear and sort of bad things. A lot of what my phone does, when you boil it down, is tell me about bad stuff, right? Like, someone was mean to me on the internet, or some terrible war has broken out, or like,

• 1:03:52Kevin Roose?

  there's a porch pirate stealing packages in my neighborhood. Like, a lot of what I get when I pick up my phone is something bad. Yeah. And so, my resolution, my goal for my tech use in 2024, is what I'm calling "more delight, less fright." Okay. So, I got this idea in part from Catherine Price, who was actually my phone detox coach back when I did a phone detox several years ago. She wrote a book about breaking up with your phone, and she actually wrote a piece recently in the New York Times about delight and the concept of bringing more delight into our lives. And she wrote that basically, you know, all these delightful things happen every day. You know, we see a pretty flower on the street, you know, a nice bird lands on a bird feeder

• 1:04:37Kevin Roose?

  outside our window, whatever delightful things, she was advocating for noticing them. And I thought, well, maybe my phone could become more delightful. Maybe if what I'm feeling when I open my phone is like a sense of dread and fear, maybe I could change that experience in some way by making my phone a more delightful place to spend time. So, I've been sort of gradually kind of rotating out some of the apps and the widgets on my phone. I took a bunch of sort of unpleasant, you know, apps that would tend to give me sort of unpleasant things the first time I opened them. I sort of put those on a second screen, and now on my home screen, it's stuff that like, makes me joyful.

• 1:05:18Kevin Roose?

  So, I made a folder in my Photos app, a new album called Delights, and I just put- Photos of things that bring me delight. You know, maybe it's my kid playing. Maybe it's a family photo. Maybe it's something that I saw on my way to the office. Like, maybe it's a screenshot from something. Maybe it's a meme that made me laugh. I just, I'm filling up this album with things that bring me delight. And I've put a little widget on my home screen that will sort of shuffle photos just from that delights album all day. So now, when I open up my phone, I get a, you know, a picture of my kid as my wallpaper. And then I open my phone and I see this little widget that has a photo of something that brings me delight.

• 1:06:00Casey Newton?

  Am I allowed to see the delight?

• 1:06:01Kevin Roose?

  You can see the delight.

• 1:06:02Casey Newton?

  Okay.

• 1:06:04Kevin Roose?

  This one is a photo of my kid at the beach. Over break, making sort of like a very joyful face.

• 1:06:14Casey Newton?

  Reaching toward the sky, that is a confirmed delight.

• 1:06:16Kevin Roose?

  Confirmed delight. You know, I'm gonna keep filling up this folder with things that bring me delight. And I just think this is like something that I am doing to try to change the emotional register with which I use my iPhone.

• 1:06:31Casey Newton?

  So I have a good sense of what's on your first screen. I would love to know, which are the sad apps that are now on the second screen?

• 1:06:38Kevin Roose?

  Well, it's everything that's sort of like, you know, like sort of work related. You know, that tends like, not a lot of times when I'm getting messages, you know, from, from a news app that, that are like, a great thing happened today. It's usually like some sort of calamity.

• 1:06:54Casey Newton?

  No one has ever experienced delight from a Slack notification either, I would say that.

• 1:06:58Kevin Roose?

  That's not true, actually. I do get some delightful Slack notifications. But, you know, I have, I've put a lot of the like, you know, the, the sort of, the stuff that just sort of makes me a little more anxious on the second screen. I actually have a red flags folder that includes things like TikTok, Threads, Blue Sky. These are not.

• 1:07:20Casey Newton?

  These are in a folder that is just marked with a red flag?

• 1:07:23Kevin Roose?

  Yeah, I'll show you. This is my, this is my red flag folder.

• 1:07:30Kevin Roose?

  But I did, I did move stuff to the first page, like the journaling app. Apple has a new journaling app. I've only just started using it, but it is helping me out. I put ChatGPT on my first screen. And I'm also putting things like, like, you know, e-reader apps to read eBooks on my screen.

• 1:07:51Casey Newton?

  Well, I think this is a great system. And there's actually only one thing that I think that would improve it, but we can actually do it right now.

• 1:07:57Kevin Roose?

  What's that?

• 1:07:57Casey Newton?

  I should take a picture of us for your delights folder.

• 1:07:59Kevin Roose?

  Aww, let's do it.

• 1:08:01Casey Newton?

  Let's do it right now. I'm just gonna take out a little phone and spin the camera, smile.

• 1:08:10Kevin Roose?

  Alright, that's going in the delights folder.

• 1:08:11Casey Newton?

  And now every time you open your phone, because hopefully you'll just set this to be the first one. You can remember when we recorded this episode. So there you go.

• 1:08:20Kevin Roose?

  I love that.

• 1:08:21Casey Newton?

  Yeah. Now, Kevin, I imagine other people might be setting their, their, their tech-related goals for the year. Do we have any tips or, or words of advice for them?

• 1:08:32Kevin Roose?

  Yeah, I, I think just be honest with yourself about what is realistic for you. I mean, one thing that has, that you've taught me about goals is that they should be something that you could actually realistically achieve. And so if the goal is, you know, never use my cellphone or, you know, never look at social media, that might not be a realistic goal for you. So I think that it should be something that is sort of a stretch, but not impossible. And I also think like, as much as you can, try not to make it, you know, try not to be too hard on yourself. Build in some buffer so that if you don't get it all the way to your goal, you still feel good about having made it part of the way there.

• 1:09:07Casey Newton?

  Yeah, I really like that. I think, you know, the one that I would just throw in there is trust your instincts. If there is a piece of software out there that is making you feel bad, just experiment with getting rid of it. You can always download it again later, right? But over and over again, when I talk to folks, they sometimes feel embarrassed because, you know, there's maybe some social app that all their friends are using, but they're not on. Trust your instinct. There is something that you know that you don't want to be a part of that, and you're probably right. And so as you're sort of casting about the tech landscape, wondering, you know, what changes you, you might want to make, I would just listen to those instincts. What do you just not want around you anymore? I promise you, you'll be able to fill it up with something you like better.

• 1:09:45Kevin Roose?

  I love that.

• 1:09:45Casey Newton?

  Yeah.

• 1:09:46Kevin Roose?

  All right, so we will check in on our goals this time next year and hopefully I will be just be full of delight.

• 1:09:52Casey Newton?

  I mean, I am excited for that and I will have found something to do besides just staring off into space while listening to Marvel Snap Game Play.

• 1:10:03Kevin Roose?

  You will no longer be the embodied version of the YouTube algorithm.

• 1:10:06Casey Newton?

  Yeah, exactly. Well, that'd be nice.

• 1:10:25Speaker 03

  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.

• 1:10:59Speaker 04

  Building and scaling AI today can feel pretty chaotic, mostly because teams are juggling too many fragmented tools at once. That's where OutSystems comes in. OutSystems gives you one unified Agentic Systems platform to manage your entire portfolio of agents, apps, and workflows. You get to innovate at the speed of AI, but with the security, governance, and control your enterprise actually needs. And the results are already there. A major manufacturer used OutSystems to build a multi-agent system that analyzes blueprints, and a global bank cut their development time by 60%. Own your Agentic future today at outsystems.com/NYT.

• 1:11:35Speaker 05

  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.

• 1:12:09Casey Newton?

  Hard Fork is produced by--

• 1:12:11Kevin Roose?

  Can you do that without jostling the thing again?

• 1:12:14Casey Newton?

  Jostling it is part of my creative process. Okay. Hard Fork is produced by Davis Land and Rachel Cohn. We had help this week from Kate Lepresty. We're edited by Jen Poyant. This episode was fact-checked by Caitlin Love. Today's show was engineered by Daniel Ramirez. Original music by Marian Lozano, Pat McCusker, Rowan Nemestow, and Dan Powell. Our audience editor is Nell Galloke. Video production by Ryan Manning and Dylan Bergerson. If you haven't already, check us out on YouTube at youtube.com/hardfork. Special thanks to Paula Schuman, Puywing Tam, and Jeffrey Miranda. You can email us at [email protected]. Let's hear those resolutions.

• 1:12:56Kevin Roose?

  And don't send us a text if you're an Android user. I really don't want to hear it.

• 1:13:00Casey Newton?

  Kevin.

• 1:13:04Kevin Roose?

  All right.

• 1:13:40Speaker 05

  You can't undo everything in life, but you can undo prediabetes. More than 2 in 5 adults in the US have prediabetes. Most don't know they have it. But with early diagnosis and action, you can delay or even prevent type 2 diabetes. Visit doihaveprediabetes.org to take the one-minute risk test and start undoing the things you can. Brought to you by the Ad Council and the Centers for Disease Control and Prevention.