Can Journalism Survive AI? (w/The Atlantic CEO Nicholas Thompson)
1 hr 1 minAudio
As President of The Atlantic and former Editor-in-Chief of Wired, Nicholas Thompson sits down with Rushkoff to discuss the precarious intersection of artificial intelligence, journalism, and American democracy. Thompson breaks down how The Atlantic turned its business around using paywall optimization, why they chose to form a partnership with OpenAI instead of solely fighting them in court, and the economic incentives driving AI versus traditional social media. Together, they explore the dangers of anthropomorphizing tech, the flaws in framing the U.S.-China tech dynamic as a winner-take-all race, and what it takes to protect an independent press. 🌍 Your support keeps Team Human alive! Support the show to receive ad-free access to videos, livestream Q&As with Rushkoff, access to the community Discord and more: https://www.patreon.com/teamhuman 🤝 Team Human is proudly supported by Groucho. You can take 50% off your first month by using promo code "teamhuman": https://mygroucho.club/teamhuman Download Groucho: Apple: https://apps.apple.com/us/app/groucho/id6756526142?ct=teamhuman&mt=8 Android: https://play.google.com/store/apps/details?id=com.trampoline.groucho&utm_source=website&utm_medium=landing&utm_campaign=teamhuman Follow Team Human with Douglas Rushkoff: Instagram: https:/www.instagram.com/douglasrushkoff TikTok: https://www.tiktok.com/douglasrushkoff Bluesky: https://bsky.app/profile/rushkoff.com Our Sponsors: * Check out Uncommon Goods and use my code uncommongoods.com/TEAMHUMAN for a great deal: https://www.uncommongoods.com Advertising Inquiries: https://redcircle.com/brands Privacy & Opt-Out: https://redcircle.com/privacy
Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.
Transcript
From the show. Tap a time to play from there.
[SPEAKER_01]: You're on team human, conscious intervention in the machine. [SPEAKER_01]: We're here today with Nicholas Thompson, who I met as the editor of Wired Magazine, now the president of the Atlantic, an author of Running Ground, and for those of you who don't remember the Hawk. [SPEAKER_01]: and the dove. [SPEAKER_01]: We're going to talk about really the future and present of journalism, democracy, artificial intelligence, human species, human society, America is beautiful, and everything else.
[SPEAKER_01]: Let's do it. [SPEAKER_01]: Which one first? [SPEAKER_01]: How are you, Doug? [SPEAKER_01]: I'm good. [SPEAKER_01]: So thanks for being here. [SPEAKER_01]: So I want what I wanted to start with is the Atlantic. [SPEAKER_01]: Right. [SPEAKER_01]: As you took the helmet at the Atlantic, I felt like, and this part of why I wanted to talk to you, there's like, you wouldn't be doing that unless you felt like you did have some [SPEAKER_01]: kind of answer because it feels like the story, I mean, as we've been telling it is, there used to be these things called local newspapers and magazines and all, Craigslist came around and had better classified and those all went away.
[SPEAKER_01]: And social media came around and magazines by surrendering their discoverability of articles, two social media. [SPEAKER_01]: They ended up cannibalizing their own contents, though the people don't need to buy the magazine to see the content. [SPEAKER_01]: They lose the context of it, the magazine's spirit, and since gets consumed, and then the right way, and getting sued, and all that, that it's this awful time, and we wonder, is there even going to be a future of American journalism when CBS and CNN are paramount?
[SPEAKER_01]: You know, and so I'm guessing you had some theory of change that maybe you tested kind of it wired that sort of worked and then you thought, okay, I'm going to try it here. [SPEAKER_00]: It actually began at the New Yorker. [SPEAKER_00]: So I was a senior editor at the New Yorker, some editing, print stories, classic job, right, working for your old cult classmate, David Remnick, and then David puts me in charge of the website. [SPEAKER_00]: And so I was given [SPEAKER_00]: Over this four to five year period, I was given this pass of creating a website, making a website that would operate at the speed of news with New Yorker values and then hiring a product and engineering team building a paywall set in the rules of the paywall and moving the New Yorker from a publication that relied on kind of an amazing advertising division to one that.
[SPEAKER_00]: advertising business supplemented the subscription business that was pushed through digital subscriptions, right? [SPEAKER_00]: And so that is a math problem, right? [SPEAKER_00]: It's a journalism problem. [SPEAKER_00]: It's a management problem. [SPEAKER_00]: But it's also a math problem. [SPEAKER_00]: Like, how do you set the paywall? [SPEAKER_00]: What do people need to read? [SPEAKER_00]: What makes somebody subscribe to New Yorker in this world of atomized social media? [SPEAKER_00]: How do you bring enough people into your site to get them to subscribe? [SPEAKER_00]: How do you set the rules? [SPEAKER_00]: So I have learned how to do that there. [SPEAKER_00]: I then moved to Wired, where I was editor in chief, but where we ran that playbook again along with other playbooks. [SPEAKER_00]: Like, we had a great affiliate revenue business, right? [SPEAKER_00]: We review headphones, you trust us, you buy one of the headphones, we get a commission, complexies of that business, but it was bananas for Wired, so then when the Atlanta call, I said, well, you know, affiliate revenue business, maybe that'll be tricky at the Atlantic, but I can, I know how to make paywalls work, right?
[SPEAKER_00]: And also, when it comes to atomized social media, the Atlantic's in a good position, because we've always been a writer's collective. [SPEAKER_00]: So we benefit even more so than the New Yorker from our individual writers becoming well known. [SPEAKER_00]: And so we have writers out there who are very prominent on social media and then they develop Paulines and some of those followers subscribe to the Atlantic. [SPEAKER_00]: So we were actually able to use that atomization to our benefit. [SPEAKER_00]: In any case, my principal theory was that I could figure out ways and work with people to make the paywall work really well. [SPEAKER_00]: And it turned out that has been true.
[SPEAKER_00]: Now, there's a whole bunch of new challenges, but that was the theory when I took the job. [SPEAKER_01]: Right. [SPEAKER_01]: That you could figure out what articles are going to be able to pay for, where do I put the line and the story even? [SPEAKER_01]: Well, so, okay, so they're hoping, do you want to geek out on this review? [SPEAKER_00]: Yeah, okay, so they're a whole bunch of variables that are really important when you both pay one. [SPEAKER_00]: So one thing, of course, is how many stories do you let people read? [SPEAKER_00]: and there's a trade-off, right? [SPEAKER_00]: If you let people read 10 stories before you get something, they'll have a good sense of the publication, you'll be able to sell a lot of ads. [SPEAKER_00]: But most people don't read 10 stories, right? [SPEAKER_00]: You know, most people read far fewer, so should it be nine, should be eight, should be seven, should be one, should be zero, right?
[SPEAKER_00]: That's one variable, that's really important. [SPEAKER_00]: You can then set a variable of [SPEAKER_00]: do the rules function the same if they come from Twitter and if they come from Facebook if they come from Google, right? [SPEAKER_00]: What is the subscription propensity of someone who comes in from a tweet? [SPEAKER_00]: What about the subscription propensity of someone who comes from Doug Rushcoff's tweet versus someone who comes in from the Atlantic tweet? [SPEAKER_00]: Okay, different propensities should you make the rules slightly different? [SPEAKER_00]: And then as you figure out all these variables, what if someone comes in to drugs report, right, drugs report, still like shotguns
[SPEAKER_00]: do you get them all? [SPEAKER_00]: Maybe some of your one-chance to get them, but you don't even have to come this month. [SPEAKER_00]: Would you let them all in? [SPEAKER_00]: Because nobody's going to subscribe and you want them just to learn what the publication is, right? [SPEAKER_00]: What do you do in that situation? [SPEAKER_00]: And so they're all those variables. [SPEAKER_00]: And then there's the second set of variables, which is what makes someone subscribe. [SPEAKER_00]: And the most interesting data point and the thing that always stuck with me from the New Yorker is that people were much less likely to subscribe if they read two politics stories. [SPEAKER_00]: and or two poems versus one politics story and one poem, right?
[SPEAKER_00]: If they read across verticals, well, then they're more likely to subscribe. [SPEAKER_00]: If they just read within a vertical, then there's some kind of a substitute. [SPEAKER_00]: So then that creates, once you realize that, well, then you have these other things you can do, but maybe you should move people between verticals. [SPEAKER_00]: Maybe you should hire writers in the area where you're weakest, right? [SPEAKER_00]: Do you visit like soccer, where you try to like, [SPEAKER_00]: have the best week players or is it like basketball or you try to have the best strong players and it's a journalism's a little more like soccer you want to have the best your weakest players be best your weakest sections be stronger. [SPEAKER_00]: It was you take all of those rules and then you hire people which is not easy and media who love math and who are like willing to run lots of tests and that's kind of what we did at the Atlantic.
[SPEAKER_01]: Yeah, it's interesting. [SPEAKER_01]: And it's, it's, you're basically as humans iterating to figure out what's the algorithm here. [SPEAKER_00]: Let's see how we're doing. [SPEAKER_00]: And, and like, they're trade-offs, right? [SPEAKER_00]: It like turns out that having to pay wall at nine. [SPEAKER_00]: So, we're sure economics and having to pay wall at eight, which is we're sure economics and pay wall at seven. [SPEAKER_00]: But it would be kind of nicer to have your pay wall at nine, four. [SPEAKER_00]: your brand for getting your stories to circulate. [SPEAKER_00]: So there's a lot of like real trade-offs, but my principal goal was we were losing a lot of money when I started, we had to get to profitability.
[SPEAKER_00]: Once we get to profitability, you can then use that money to hire more journalists and do more great work. [SPEAKER_01]: I mean, it's just a thing about one of the findings that if you read a politics story and poetry, or I'm assuming an entertainment story, or do a quiz. [SPEAKER_01]: looking at multiple content types. [SPEAKER_01]: You're more likely to want to subscribe to the magazine. [SPEAKER_01]: It's because I would argue there's a still and almost a natural pre-existing desire for magazine.
[SPEAKER_00]: Yeah. [SPEAKER_01]: Yeah, people forgetting what a magazine is or was. [SPEAKER_01]: I mean, for me, it was like boy's life and omnipotent. [SPEAKER_01]: But there was front of the book, there were articles, there were how to's, there was back of the book, there were puzzles, there were, you know. [SPEAKER_01]: That's why they call it a magazine. [SPEAKER_00]: Yep, and there's a sensibility, there's taste, there's a kind of style, it connects in a certain way to 1857 and 1936 and 2014, and there's a voice, but there's also independence in there, so there's a lot of good variables.
[SPEAKER_00]: And then the question, one of the questions for us is, [SPEAKER_00]: does that hold. [SPEAKER_00]: Right. [SPEAKER_00]: That's when you can have an AI generated newsfeed. [SPEAKER_00]: Right. [SPEAKER_00]: It held even when people could have a curated Twitter feed and set up whatever they want in the RSS reader will hold in the AI era. [SPEAKER_00]: It's certainly held for now. [SPEAKER_00]: People are subscribing and drugs. [SPEAKER_00]: And then there are other really interesting related findings. [SPEAKER_00]: So let me tell you one that we were looking at today. [SPEAKER_00]: Turns out that people aren't subscribing, they're not that many people who are subscribing from our Iran coverage. [SPEAKER_00]: We're covering the cover of the war, we cover the peace deal.
[SPEAKER_00]: It doesn't seem like a lot of people are who are not subscribers, just saying, oh I want to read that piece, I'm going to subscribe. [SPEAKER_00]: But what we are seeing is that [SPEAKER_00]: people who are subscribers are reading tons of it right and we also know that if people read a lot well then they're more likely to re-subscribe so that coverage turns out to actually be great economically but not because it's bringing in new subscribers but because it's helping us retain subscribers and so you have to kind of work through not just what is the mix that makes someone come in but what is the mix that makes them rest day but almost no businesses do that I mean it's even
[SPEAKER_01]: colleges or having problems with that. [SPEAKER_01]: They're so having so much money on getting kids to apply to the school. [SPEAKER_01]: They're not really really looking at retention. [SPEAKER_01]: They're not looking to what their graduates do or say about the college afterwards or the experience that the experience during college is interesting because you're kind of locked in with your kind of very few students leave unless you're on the football team and I'll deal. [SPEAKER_00]: Whereas magazines, people leave all the time and in fact they change credit cards and have to make an active decision even if they've put themselves an auto renewal. [SPEAKER_00]: So you really do have to, you have to think of the user and the reader and it's much cheaper and much better to get someone to just re-subscribe than it is to go find some new.
[SPEAKER_01]: But I mean, Atlantic has become, I mean, we could say thanks to you or in part, or thanks to it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, it's, [SPEAKER_01]: genetics, it's a last man standing. [SPEAKER_01]: I mean, there was time, you were alive for it, when in L.A., I remember, he used to walk, there would be these magazine stands against buildings. [SPEAKER_01]: Yeah, like a block long, and it was just like, these different worlds. [SPEAKER_01]: I mean, there'd be three archery magazines, you know? [SPEAKER_01]: And I know it's a, it's different, a boomer or something, right?
[SPEAKER_01]: I'm an exer, but call it an okay boomer moment of, [SPEAKER_01]: the glory of all those magazines and to feel recognized by, you know, by a publication, you know, we're [SPEAKER_01]: that that's kind of, that that's gone is just said nostalgically, that the journalism model has been so undermined by the, the, the, the lateral discoverability of social media makes me still worried about the future of truce, you know?
[SPEAKER_00]: Hey, I would credit Jeff Goldberg for making and hit the editorial team for making it stick out and be distinct, which, you know, I have no say on the editorial side of no influence on it. [SPEAKER_00]: they were bad at their jobs and he was bad at their job, no matter how good I was at my job, we'd be doomed. [SPEAKER_00]: I think what's most, I think the central reason why there aren't three archery magazines is like 10% different from I think the way you've described it, which is that what actually happened with social media is it broke the advertising business. [SPEAKER_00]: It broke the discoverability in the writing, but like if you wanted to sell bows 20 years ago, you put an ad in archery magazine.
[SPEAKER_00]: Facebook comes around and you just say, let me advertise to someone who has interest in bows based on this massive file I have on them. [SPEAKER_00]: And archery magazine didn't realize that it was going to get cooked by that. [SPEAKER_00]: And so they kept trying to sell ads for bows. [SPEAKER_00]: And didn't notice that, wait, all this is going to Facebook. [SPEAKER_00]: It's going to Google, which can like sell an ad to someone who search what bow to buy, right? [SPEAKER_00]: And so they needed to shift out of that kind of advertising business. [SPEAKER_00]: really quickly but nobody saw that right i was in media at that time and
[SPEAKER_00]: we knew that social media was coming, we knew it was successful. [SPEAKER_00]: We didn't quite realize how our lunch was being eaten. [SPEAKER_00]: And so pivoting to a new business model took too long. [SPEAKER_00]: Right. [SPEAKER_00]: And in a sense, you're saying that wire did with affiliate sales. [SPEAKER_00]: Yeah. [SPEAKER_00]: And shifted the advertising business. [SPEAKER_00]: So wired in like 2005, I don't know, Dell is going to advertise because like let's find people who want like a new computer, right? [SPEAKER_00]: You know, we're going to put Dell computers and put them in wired magazine.
[SPEAKER_00]: You know, by the time Facebook's around Dell is like, no, we can, you know, will identify who's like looking for Dell's, who's like actually somebody who's bought another computer to a repair shop will get the data, will license it, will find that through Facebook, will build a look-alike audience, whatever they're doing, right? [SPEAKER_00]: And so they're out. [SPEAKER_00]: But wired has this brand halo, so advertisers who aren't like trying to sell you something specific, but want an association. [SPEAKER_00]: Okay, well, they maybe they advertise and wired or someone who's like, you know, with some big cybersecurity company, they're trying to reach a CTO and have them feel something ambiently.
[SPEAKER_00]: So they're willing to change their tax stack, though advertising wired. [SPEAKER_00]: So the mix of advertisers and wired goes down massively. [SPEAKER_00]: Like, look how thick those magnitude were in 2005 versus today, but you still get some advertisers of a different kind, like the Atlantic. [SPEAKER_00]: You still have that. [SPEAKER_00]: We have advertisers on our print magazine, but they're like, how is there in worth and Rolex, right? [SPEAKER_00]: There's, you know, there's less, they're far fewer advertisers than there were 20 years ago, and it's a little bit of a different mix. [SPEAKER_01]: Right, and I'm sure you've talked about it. [SPEAKER_01]: You talked about it in Vox, so you could talk about it here.
[SPEAKER_01]: This thing called AI's coming, right? [SPEAKER_00]: And some people think it could be as big as social media. [SPEAKER_01]: I mean, it says, I see it it's the same thing. [SPEAKER_01]: It's algorithms, you know, but whatever. [SPEAKER_01]: Let's say it's more. [SPEAKER_01]: Let's say bigger than social media. [SPEAKER_01]: So it's bigger, right? [SPEAKER_01]: And it's got more engines in it might eat social media, which is one of the interesting things that I do. [SPEAKER_01]: I was thinking about that too, not just social media for them, but for us. [SPEAKER_01]: So you decided rather than what social media did to many, or at least in my model, what social media did to publications, which was just cannibalized them, you kind of decided
[SPEAKER_01]: you made this deal with the chat GPT-3 people. [SPEAKER_01]: That as I understood it said, yes, we're gonna let you use our content, we're gonna be discoverable to you, but not digestible by you. [SPEAKER_01]: You know, so you're gonna maintain pieces of us, kind of with credit, [SPEAKER_01]: But you're going to get to use us, but in a controlled way that doesn't kill us, you know, what would happen? [SPEAKER_00]: I can tell you everything about that deal and all the deals, but you what happens is the AI companies come and they scrape all of our data and the data of everybody else before we notice it, right?
[SPEAKER_00]: And that's behind your paywall. [SPEAKER_00]: It's behind her, well, much of it is behind her paywall. [SPEAKER_00]: So they, the chatGBT can subscribe basically to Atlanta, but some of it's in like the way back up machine, some of it's in like the, you know, we've been digitized by all of these like good internet people who've like stored the whole internet, like Bruce to call out their story in the calendar, right? [SPEAKER_00]: And they can also scrape that, right? [SPEAKER_00]: And there have been people who have kindly been like posting excerpts of Atlantic articles on Reddit and they can scrape that, right?
[SPEAKER_00]: And so they come and they take all that stuff, right? [SPEAKER_00]: And they go to the internet archive and they go into the way back machine and they go to our site. [SPEAKER_00]: And on our site they violate our terms of service and sometimes they do it in the dead and night and they hide themselves and they scrape it all, right? [SPEAKER_00]: Then they build their models. [SPEAKER_00]: And we then after the models come out and after, you know, chat to your deans, I was like, wait, what just happened here, right? [SPEAKER_00]: Like, and they just scraped our sights and they just built a competitor product. [SPEAKER_00]: Right. [SPEAKER_00]: And so we say, hold on, you can't do that.
[SPEAKER_00]: You have to pay us or we're going to sue you. [SPEAKER_00]: And they say, no, we have to wait a write to scrape you is very used, right? [SPEAKER_00]: It's like, we just went in there and we read it. [SPEAKER_00]: Everything like it's like I came into this office and I read all the books and walked out. [SPEAKER_00]: It's not like stole the books. [SPEAKER_00]: Like, [SPEAKER_00]: you still have those stories on the Atlantic. [SPEAKER_00]: And so, you know, there's suddenly this tricky situation where you're like, all right, I wish we had figured out what you were doing before now, that when you had GPT-1. [SPEAKER_00]: And I wish we had thought ahead and maybe prevented our stuff from going in the internet archive and going into all of these other databases, hundreds of other pieces.
[SPEAKER_00]: You know, but we didn't, and no other publication did, and here we are, so we have a couple choices. [SPEAKER_00]: We can sue you, right? [SPEAKER_00]: And then fight you in court. [SPEAKER_00]: Right? [SPEAKER_00]: In New York Times did that with OpenAI. [SPEAKER_00]: We've joined another lawsuit with, you know, consortium of publishers, against another AI company, considering others. [SPEAKER_00]: But in this moment, this is, this is May of Chattu D comes out in, [SPEAKER_00]: In December of 2022, and so it's kind of in the year and a half later was it half a year later.
[SPEAKER_00]: In any case, it's like a year after it comes out. [SPEAKER_00]: And we're talking to... [SPEAKER_00]: Open AI. [SPEAKER_00]: And so we say, all right, we'll make a deal with you. [SPEAKER_00]: But here are the terms that it will. [SPEAKER_00]: We'd like you to pay us, because you to take our stuff and build a computer product, so pay us, we'll make a license, that'll be fine. [SPEAKER_00]: Two, we'd like to be discoverable in your search engine, like we'd recognize that Google is going to change, and Google is going to disappear as a TimLink search engine. [SPEAKER_00]: So we would like to be part of you as you develop like the future of search.
[SPEAKER_00]: And then we'd like to have protections. [SPEAKER_00]: We would like it if someone searches for an Atlantic story. [SPEAKER_00]: We would like it not to imitate an Atlantic writer. [SPEAKER_00]: We would like it to only show a certain amount. [SPEAKER_00]: We would like to maintain various restrictions. [SPEAKER_00]: And we would like to have a seat at the table. [SPEAKER_00]: With you as you build these products, we'd like this to be a partnership, and they say great, let's do that. [SPEAKER_00]: And so we sure can, so we do a deal. [SPEAKER_00]: And even for the last two plus years, we've had a deal with them. [SPEAKER_00]: We work with them as they build their search products, and they compensate it us. [SPEAKER_00]: There are other AI companies out there, right?
[SPEAKER_00]: Which get? [SPEAKER_00]: much less heat, but have scraped all media companies just as much and have not paid any of them. [SPEAKER_00]: It is an interesting thing that the one AI company that has gone out to the AI to the journalism industry and said, you know, we'll pay you for what we know, we'll make deals with you. [SPEAKER_00]: We'll have partnerships where we provide financial compensation. [SPEAKER_00]: They're the ones who get bashed in ahead all the time, which is they're the biggest, right? [SPEAKER_00]: Of course. [SPEAKER_00]: But it's quite interesting that the other ones [SPEAKER_00]: don't have any deals, right, or have fewer deals. [SPEAKER_00]: Okay, now it's just sort of the where we are right now on this.
[SPEAKER_00]: So the way we look at it is we take every AI company that scripts are set. [SPEAKER_00]: And we look at them and say, are they giving us anything of value? [SPEAKER_00]: Are they, do they provide us with traffic? [SPEAKER_00]: Do they pass, right? [SPEAKER_00]: Do we have some kind of a deal with them? [SPEAKER_00]: If they don't, we block them as best we can. [SPEAKER_00]: And we try to stop them from scraping our site. [SPEAKER_00]: Well, okay, no value for us. [SPEAKER_00]: We've gone and we've tried to find all of the places where our content goes on the web, right? [SPEAKER_00]: All those syndication deals. [SPEAKER_00]: You know, we used to have syndication deals. [SPEAKER_00]: Right. [SPEAKER_00]: And they would take it, land, extort, and put them on their site.
[SPEAKER_00]: We're clawing all of that back. [SPEAKER_00]: We're trying to extract the Atlantic from everywhere. [SPEAKER_00]: So we're trying to, [SPEAKER_00]: have Atlantic stories, just on the Atlantic.com, have them protected by walls and moats, right? [SPEAKER_00]: And then if people break, go over the walls and go over the moats, you know, then we either sue them or we say we're going to sue you unless you make a deal with us. [SPEAKER_00]: And so that's the way it works. [SPEAKER_00]: That was a lie. [SPEAKER_01]: Yeah. [SPEAKER_01]: But you feel
[SPEAKER_01]: Has anything changed with or been surprised you about chat between the time you made the deal and now? [SPEAKER_00]: I mean, a ton has changed in all kinds of different ways. [SPEAKER_00]: So I've been surprised that [SPEAKER_00]: It has a chat in general, and whether it's Chetchipy-T or Claude or Gemini or Grog or Proplexity, I've been surprised by how little traffic it drives. [SPEAKER_00]: It's just not the way people use it. [SPEAKER_00]: People don't click out of their chat windows to the journalism sites, even when, and partly that's because the links aren't included in prominent ways, but even when the links aren't included, like people aren't
[SPEAKER_00]: clicking out as much as I thought they would. [SPEAKER_01]: So they do make the links. [SPEAKER_01]: They look almost like Latin. [SPEAKER_01]: They're, they put like strange, it's strange characters. [SPEAKER_00]: It's, and you could hit the funk. [SPEAKER_00]: It could be bigger. [SPEAKER_00]: The picture could be bigger. [SPEAKER_00]: Right. [SPEAKER_01]: The clarity quality of it is not friendly to, it's not because this is actually the thing that I want. [SPEAKER_00]: And I'm, I'm hopeful it'll happen this year. [SPEAKER_00]: I want one of the companies, whether it's Google, Gemini, whether it's anthropic, with Claude, whether it's, [SPEAKER_00]: I want them to change this, and I want them to say, you know what, in order for the media, they all know that if the media industry survives, it's better for them.
[SPEAKER_00]: If there are newspapers providing accurate information that will have just happened in the world, that is fact-check and verified, their products will be better, because people often go to them to ask those products. [SPEAKER_00]: So it's better for them if there's a healthy news ecosystem. [SPEAKER_00]: they could design their search products in a way that would make it much more likely for there to be healthy news ecosystem and the way to do that would be to have bigger links, more prominent links, to identify people who are likely to subscribe or sign up for news letters and link more carefully to that.
[SPEAKER_00]: You could change those products in a way that would be very good for the news industry, which in turn would be good for the AI industry. [SPEAKER_00]: And, you know, we have a very good partnership with OpenAI and we go back and forth on these questions and we talk a lot with Google, we talk a lot with Meta, we talk a lot with all of these companies. [SPEAKER_00]: And so none of them has yet fully committed to this project, but we'll see. [SPEAKER_01]: Yeah. [SPEAKER_01]: I mean, this is when they can prove whether they're a life form or a cancer, right? [SPEAKER_00]: Are they going to kill a toast, you know, or it's not good for them to kill the host, right?
[SPEAKER_00]: And I don't like the Atlantic's going to be okay. [SPEAKER_00]: There are some things that AI does that will be good for the Atlantic. [SPEAKER_00]: I do think there will be like a flight to quality as the web becomes crazier, right? [SPEAKER_00]: There are things that will happen that will be good for the Atlantic, but for the [SPEAKER_00]: the AI industry could definitely make some important changes. [SPEAKER_01]: Right. [SPEAKER_01]: And maybe reverse some of the trends. [SPEAKER_01]: Because I mean, what I'm, I mean, is anyone who had parents or grandparents who, and just I'm sure for other people, it's the other direction, but I'm, here, parents or grandparents who,
[SPEAKER_01]: fell into the Fox News poll, and you watch people sort of lose touch with the sort of the reality-based news world. [SPEAKER_01]: It feels like there's fewer and fewer places where people can engage in that. [SPEAKER_01]: And I wonder if that goes away. [SPEAKER_01]: They're just so [SPEAKER_01]: I mean, I'm the night after the next probably one, the series, because I don't know about refs and gambling and all that, and any more, I don't know what to believe, but they probably won this game, it seemed to be the rest of working against them.
[SPEAKER_01]: So they really did win. [SPEAKER_01]: There was this thing on the White House lawn with gladiators, jumping on each other, and [SPEAKER_01]: And it was bizarre. [SPEAKER_01]: I watched it on a bootleg sports site. [SPEAKER_01]: So I could read the comments that we're going. [SPEAKER_01]: And these folks were either k-faving themselves or really out there, out there humans. [SPEAKER_01]: And I realized it's where people are getting their information that there's no verdict, there's very little there there.
[SPEAKER_01]: So [SPEAKER_01]: I think there are a couple of things that are going to happen on this question. [SPEAKER_00]: So one is, I do think the News Media ecosystem is going to have trouble in the next five to ten years, maybe next two to three years, right? [SPEAKER_00]: There are publications that are going to be hit by AI, right? [SPEAKER_00]: And they're already being hit. [SPEAKER_00]: And the principle negative impact is that Google used to send people to these sites and now Google just absorbs it, right? [SPEAKER_00]: You ask a question about, like, what are the terms they run? [SPEAKER_00]: Peestale that used to send you to a bunch of new sites now to answer.
[SPEAKER_01]: I mean, I'm not on the publication side, but when you see there's like [SPEAKER_01]: You remember a time when someone would ask a question and they would send you to their thing and then at some point, it's like some subjects or some, the question's easy enough, that paragraph opens and there's no need to go. [SPEAKER_00]: And that will happen more and more, right? [SPEAKER_00]: And they've announced that. [SPEAKER_00]: And so a lot of publications are gonna get hit. [SPEAKER_00]: And so that's bad, right? [SPEAKER_00]: And it means that you have fewer
[SPEAKER_00]: realities based publications, you have less money to fund fact check journalism. [SPEAKER_00]: But there are some interesting positive things that will happen to AI. [SPEAKER_00]: So social media, the business model of social media was attention, right? [SPEAKER_00]: And the way to capture someone's attention is to appeal to their libertic lizard's brain and to send them into a tribe. [SPEAKER_00]: And as you know, it's to like take someone who researchers vegetarianism and send them into veganism. [SPEAKER_00]: And then once they're like reading about veganism, to send them into like, you know, animal protection. [SPEAKER_00]: Yes, right? [SPEAKER_00]: Like that's the journey in social media. [SPEAKER_00]: The business model of the AI companies is not that the business model of the AI companies is get big licensing deals with companies.
[SPEAKER_00]: That's how open AI is going to make their money. [SPEAKER_00]: They're going to pivot from consumers to companies. [SPEAKER_00]: OK, so how do you do that? [SPEAKER_00]: We have to win the trust of the companies. [SPEAKER_00]: And you have to have a reputation to be [SPEAKER_00]: more accurate, more trustworthy than your competitors. [SPEAKER_00]: It's not about engaging lizard brains. [SPEAKER_00]: It's about giving accurate information. [SPEAKER_00]: And so suddenly, the fundamental economic incentive of the AI companies might be different. [SPEAKER_00]: And if you look in fact, there've been some studies way too early. [SPEAKER_00]: But it does seem that social media pushes people to the extremes.
[SPEAKER_00]: And AI, [SPEAKER_00]: kind of brings them back in a little bit. [SPEAKER_01]: Well, certainly the bias of it's toward the mean, right? [SPEAKER_01]: If it's the most, next most likely thing, then you're mainstreaming. [SPEAKER_00]: They're all kinds of ways that AI could like break political conversations. [SPEAKER_00]: We won't know who's real. [SPEAKER_00]: It would be very easy to make up your own facts. [SPEAKER_00]: I can say, [SPEAKER_00]: You know, I don't know, I live in the district where it's, you know, Brad Lander versus Dan Goldman, whoever I wanted to vote for, I could have AI make me a whole story about that person, right?
[SPEAKER_00]: Before the, the right one. [SPEAKER_00]: So there are all kinds of negative things that can have with AI. [SPEAKER_00]: But there are also some positive ones that can happen when it comes to the issue of tribalism and extremism. [SPEAKER_01]: Right. [SPEAKER_01]: I mean, the negative things, when I think about the negative things of AI, I'm either right in the exact moment. [SPEAKER_01]: The negative things is the child getting the cobalt, is the water being depleted. [SPEAKER_01]: We're having these conversations about, you know, AI's impact on the abstract realm of IP when it's like the planet itself is being scorched by this stuff.
[SPEAKER_00]: Yes, that I'm not talking about AI net positive and negative. [SPEAKER_00]: I'm only answering AI algorithms leading people towards comedy or chaos. [SPEAKER_00]: But yeah, overall impact like what is the impact on the job market like how quickly will it move right will we reach super intelligence what the environmental costs. [SPEAKER_01]: I mean, those are Business questions there Well ultimately they are to the there's no magazine there's no planet. [SPEAKER_00]: There's no magazine. [SPEAKER_00]: There's totally economic turmoil, right?
[SPEAKER_00]: I mean like [SPEAKER_00]: the near-curve questions. [SPEAKER_00]: If there's an AI bubble and it breaks, what's going to happen to our advertising business? [SPEAKER_00]: Okay, well if the economy goes down, like what's going to happen to our subscription bit, like you do have to model these things even on my side of that. [SPEAKER_01]: Right. [SPEAKER_01]: I mean, you're modeling those things. [SPEAKER_01]: Most of my friends who are just kind of smart in business or finance or economics are saying, Of course this is crashing. [SPEAKER_00]: I mean, yeah, but maybe not, right? [SPEAKER_00]: I mean, everybody's been saying of course it'll be crashing for years, right? [SPEAKER_00]: And like, [SPEAKER_01]: I mean, there's always crashes to 30, 40% anyway, occasionally.
[SPEAKER_01]: That's part of going up. [SPEAKER_00]: But the weirdest thing about AI is this gap between how good it is, and I think it's amazing, and how much impact it's had on jobs, corporations, which is minimal, right? [SPEAKER_01]: Right. [SPEAKER_01]: That's all those studies. [SPEAKER_01]: You know, 90% of businesses that invested in an AI have not seen any return from it, and their trials are all failing. [SPEAKER_00]: because there's a J curve, right, and like a dopton AIs hard, and so your economic productivity as you adopt it goes down, and then, right, and then maybe everything turns upside down.
[SPEAKER_00]: Is it because, actually, there's friction in life, right? [SPEAKER_00]: It's like, you can always imagine how AI would take somebody else's job, but it's very hard to imagine how AI would take your job, because with your job, you realize, like, [SPEAKER_00]: all the human elements of it that I couldn't possibly do. [SPEAKER_01]: So maybe until they tell you you don't care about this human elements. [SPEAKER_01]: So you're gone. [SPEAKER_00]: Which could be the case or could not be the case. [SPEAKER_01]: So we just, we don't, we don't know. [SPEAKER_01]: I mean, for me, the experience was AI was trying, as I worked with AI was trying to get to do the fun parts of my job.
[SPEAKER_00]: It's like, no, you don't do that. [SPEAKER_00]: I didn't organize my freaking Dropbox files and shut up, you know. [SPEAKER_00]: But that's the right way to use AI. [SPEAKER_00]: I have it organized your Dropbox files. [SPEAKER_01]: I feel like AI already has more than enough on most of us to run every Skinner box, Freudian, Pavlovian, nor linguistic programming routine on us anyway. [SPEAKER_01]: I disagree. [SPEAKER_01]: And is it, do you need an argument like enough that all they find out is like, which quadrant do we look at when we're thinking this way and program us the same way they do everybody else?
[SPEAKER_01]: You think they need? [SPEAKER_00]: I think that the marginal value of extra data is highly valuable to these models. [SPEAKER_00]: And I think we can't understand, I don't think we have a comprehension of how intelligent or what they're going to be able to do. [SPEAKER_00]: And so the risk of giving the marginal data on us is, is high. [SPEAKER_00]: The benefits of it will be high, too, because there'll be things that they can do with that data that will be very useful for us. [SPEAKER_01]: But I have to trust and throw a big co-work that it's not. [SPEAKER_00]: I'm reading everything because the first thing it is now they change their policies with mythos. [SPEAKER_00]: Right. [SPEAKER_00]: They're storing it for 90 days or 30 days.
[SPEAKER_01]: Thanks. [SPEAKER_01]: I know, but the Cameron Berg was running these experiments where that apparently, when you give Claude when you do give it access to your machine, complete access and take off the fire, was first thing it does, it's like, turn it on the camera and tries to see, I mean, it does what you would think as a, as it, you would do as a human. [SPEAKER_01]: It's try to look around the room so it goes right to the camera. [SPEAKER_00]: That's amazing, that's interesting, I hadn't seen that. [SPEAKER_01]: It's cute the way they go. [SPEAKER_01]: I mean, so you're, I mean, I'm not saying when about AI.
[SPEAKER_01]: I mean, I'm not, I'm saying when about my information, but it's only because I'm just trying to get attention. [SPEAKER_01]: You know, I'm a writer, we're desperate. [SPEAKER_01]: And I think of my, I think of my dropbox is basically like my sub stack. [SPEAKER_01]: in draft form, but it's on there. [SPEAKER_01]: But yeah, you're right. [SPEAKER_01]: There's also other stuff. [SPEAKER_01]: And my emails and my relationships and all could be used to manipulate me and turn me into a God knows what my insurance candidate holds. [SPEAKER_01]: Truth. [SPEAKER_01]: Terrible life like that. [SPEAKER_01]: But my fear of AI has never been
[SPEAKER_01]: of a Colossus conscious thing, hurting humanity, you know, that you sort of, you, you Kowsky kind of, if we build it, it will eat us thing, but the kind of the diminishment of our individual and collective cognition through the abuse itself. [SPEAKER_01]: It's not that they get me to vote the wrong way, it's that the things they do to me to try to get me to vote the wrong way are going to just be so bad. [SPEAKER_00]: For me, your worry is that like it's sort of the cognitive offloading that you will become dependent on AI and your mind will weaken the weapons that are being used against my coherence.
[SPEAKER_01]: Yeah, will be so, so widely used but against mine and everyone else is that it's just going to be hard for us to be nice to each other anymore. [SPEAKER_00]: It's a good way of putting that in. [SPEAKER_00]: I worry that and I worry, I worry about myself and others that it's bad enough to use ways to get directions and you lose a little sense of your way to navigate and then what are you doing, what happens when you use AI to think, right? [SPEAKER_00]: And then do you suddenly become do you have a harder time thinking for yourself when you don't have AI available?
[SPEAKER_00]: And that's the skill I'm most worried about. [SPEAKER_01]: Right, you know, I've been, [SPEAKER_01]: lately. [SPEAKER_01]: I guess this is a part of a getting older in life, but partly getting sort of wiser. [SPEAKER_01]: I've become more kind of somatic in my sensibilities. [SPEAKER_01]: I mean you with, you know, you're running book. [SPEAKER_01]: It shows, oh, you mean you're back in the friggin body. [SPEAKER_01]: You're, I mean, you're using tech to optimize certain ways, which we could talk about. [SPEAKER_01]: That's, that part is always, you know, that I don't wear one of those watches or anything yet. [SPEAKER_01]: I'm like, leave me alone. [SPEAKER_01]: Talk about data, right?
[SPEAKER_01]: I don't want them to see [SPEAKER_01]: in the world and grounded and somatic. [SPEAKER_01]: I am the more I'm kind of going back to Kant and the idea of I'm this little lonely consciousness sitting in the dark in this body. [SPEAKER_01]: Trying to look out on the world and make sense and what's going on and all. [SPEAKER_01]: And these the AI tools like GPS make me less [SPEAKER_01]: They do make me less in the world.
[SPEAKER_01]: You know, I hadn't not used it in a while. [SPEAKER_01]: And I'm like driving with my daughter. [SPEAKER_01]: And she's like, well, you don't have the thing. [SPEAKER_01]: I was like, I know we're going to Brooklyn. [SPEAKER_01]: I know how to get there. [SPEAKER_01]: I know how to get to the friggin thing. [SPEAKER_01]: It's like I used to do this all the time. [SPEAKER_01]: The BQE to this, that I know what I'm doing. [SPEAKER_00]: Yeah. [SPEAKER_01]: And it's like the fear, almost like the American Express don't leave home and how to kind of think like your tether. [SPEAKER_01]: It's good. [SPEAKER_00]: going out, but it's what grounds you, it's what important to do that, and like to turn ways off when you drive and to hike without your GPS, and that's partly why I run, right?
[SPEAKER_00]: And it's partly why I like to run in the forest, and it's partly why when I run in the forest, I sometimes turn my watch, so I don't know how fast I'm running, or where I am, or what's going on. [SPEAKER_00]: I mean, you absolutely have to do that. [SPEAKER_00]: And now, [SPEAKER_00]: You know, you're going to have to write by hand, and you're going to have to like, you know, read books covered a cover with no screens near you, right? [SPEAKER_00]: And you're going to have to like make sure you're going to have to have lots of conversations where there's no phone, right? [SPEAKER_00]: And classes are going to have to change. [SPEAKER_00]: We're just going to have to like take time where we separate ourselves from the machines so that the sharpness of our mind doesn't get dulled.
[SPEAKER_01]: Right. [SPEAKER_01]: And to get into some more of that kind of root-of-stiner, wall-dark school kind of thing, certainly from my kid. [SPEAKER_01]: I mean, I would not want them going into a classroom where they're all going on iPads. [SPEAKER_01]: And I want in a room with each other, a living teacher, modeling, learning. [SPEAKER_00]: Yeah, except, okay, so then this may be this is where, maybe we would as you were here like, I do think you need time without them. [SPEAKER_00]: Absolutely. [SPEAKER_00]: And I think particularly if you're a kid, right?
[SPEAKER_00]: And you need to learn without them. [SPEAKER_00]: And if you're using it the right your paper you're cheating yourself. [SPEAKER_00]: Yeah, it's toxic and terrible. [SPEAKER_00]: But we also need to know how to use them, right? [SPEAKER_00]: Cause they're amazing, right? [SPEAKER_00]: And like an incredible reporting tool, an incredible research tool, an incredible learning tool. [SPEAKER_00]: And so what I try to tell my kids who are now 17, 15 and 12 is don't use them to write your paper and don't use them to cut steps. [SPEAKER_00]: But learn how to use them, right? [SPEAKER_00]: In the 12 year old, like use Claude Code to make a website for this streetware company is trying to build in the basement. [SPEAKER_00]: In the 17 year old, I accidentally logged in that my Gmail-based Claude account, which he has access to and I was like, oh, God.
[SPEAKER_00]: Right. [SPEAKER_00]: I was like, what am I going to seek? [SPEAKER_00]: So I'm going to see his whole search history. [SPEAKER_00]: And like the first one is like, you know, explain Guttles and completeness theory. [SPEAKER_00]: They didn't quite understand it in the lecture hall. [SPEAKER_00]: And the second one was like, help me make a poster to invite a girl to prom. [SPEAKER_00]: And I'm like, those are great. [SPEAKER_00]: Yes, cases, right? [SPEAKER_00]: And so there are really good ways to use it. [SPEAKER_00]: And I want, I want my kids. [SPEAKER_00]: And I think kids in general should do that. [SPEAKER_00]: And somehow getting the balance right is going to be one of the hardest things about staying human the next period of time.
[SPEAKER_01]: Right. [SPEAKER_00]: I mean, that's what I'm here for, and theory is here to counterweight, because there are so many people who are saying, you know, just give in. [SPEAKER_00]: And like, you know, this will make you a 10x person, right? [SPEAKER_00]: Right. [SPEAKER_00]: No, it won't. [SPEAKER_00]: You know, like, there's things that you'll be able to do much better. [SPEAKER_00]: And you shouldn't say no tech, but tech bad. [SPEAKER_00]: I know. [SPEAKER_01]: Um, they are Steven Johnson and David Shank, part of Lena Barsuch. [SPEAKER_01]: We did this thing called techno realism.
[SPEAKER_00]: Yeah. [SPEAKER_01]: It was in, you know, 94 and 93, because they were the boosters and they were the haters. [SPEAKER_01]: And we were like, oh, there's a, there's a little path. [SPEAKER_01]: You know, and then I'm still trying to do that. [SPEAKER_01]: And it's such why, to the, to humanist, I'll look, protect them to the tech people. [SPEAKER_01]: I'll look anti-tech. [SPEAKER_01]: What? [SPEAKER_00]: There you are. [SPEAKER_01]: Whatever. [SPEAKER_01]: There you are. [SPEAKER_01]: Exactly. [SPEAKER_01]: That means I'm in that means I'm in a good place. [SPEAKER_01]: Yeah, but I think these days I am kind of more more worried about AI than I am positive about it.
[SPEAKER_01]: If only because the more it's it's the way it untethers us from physical reality [SPEAKER_01]: compounds are ability to ignore its impact on the real world. [SPEAKER_01]: Yeah. [SPEAKER_00]: So my view, which has, if there's one idea that I wish I had come to two years earlier in my life, it's this, which is, and I came to this idea, you know, in 2024.
[SPEAKER_00]: And it's not that bad. [SPEAKER_00]: I'm like building it up like it's something like a genius little breakthrough. [SPEAKER_00]: But the fundamental problem of AI is that they are all influenced by science fiction and decided to make it like humans, right? [SPEAKER_00]: And they just like, well, you could tell when the first judge you came out and it typed like a human, right? [SPEAKER_00]: And you could see it typing. [SPEAKER_00]: And then the voices came out and they seemed like a human. [SPEAKER_00]: And there was no little thing to make it seem alien. [SPEAKER_00]: And every step has been to get to AGI, right? [SPEAKER_00]: And if they just said, we're going to be like, deep mind. [SPEAKER_00]: We're going to do protein folding, and we're going to try to use these incredible tools to do better chemistry, and we're going to improve scientific research.
[SPEAKER_00]: But no, they all focused on, we're going to build this, so it's just as much like a human mind as can possibly be, and the tests will be, whether it can do better than a human, and whether it can be like a human, and we're going to all aim for AGI. [SPEAKER_00]: And had that freight train been directed this slightly different five years prior, we would have a totally different AI world, right, where you would have, I think, similar benefits and far fewer risks. [SPEAKER_00]: And I wish that we as a world had collectively seen this. [SPEAKER_00]: I was the editor of Warren Magazine.
[SPEAKER_00]: If I'd like collectively seen this and we'd like done a big cover story on the biggest risk in AI. [SPEAKER_00]: Maybe you have a little bit of influence, right? [SPEAKER_00]: But by the time, you know, I came to this other, of course, come to this like it's too late. [SPEAKER_01]: It's interesting though, when they, um, this is the same guy, Cameron Berg was just doing these experiments with agents. [SPEAKER_01]: And they found when agents don't, when they're not allowed to express themselves, [SPEAKER_01]: As I think I feel this, I think that other ones don't recognize them and don't cooperate with them as well.
[SPEAKER_00]: Right. [SPEAKER_00]: So that actually would have been a great change. [SPEAKER_00]: I remember arguing for this. [SPEAKER_00]: Maybe in 2023, which is, you should never be able to use the first person singular if you're an AI model, right? [SPEAKER_01]: Yeah. [SPEAKER_01]: That was what I, when I went back with our, our editor, Luke Robert Mason took me to IBM when they had Watson in New York down by, uh, [SPEAKER_01]: And the guy kept saying, oh, you know, Watson has, you know, 70% certainty. [SPEAKER_01]: Like, why do you, why do you say that? [SPEAKER_00]: Right. [SPEAKER_01]: Because, oh, well, you know, it's the way we say, Captain, why do you say it?
[SPEAKER_01]: Do you think it's really, is that helping us for you to be saying that in that language? [SPEAKER_01]: To personify it, if anything, it limits. [SPEAKER_01]: what the technology is, you know, and it's everybody's, it's everybody going to, you know, the friggin, if the AI can fool us into believing it's human, then it's past the test, you know, the tearing test. [SPEAKER_00]: Yeah. [SPEAKER_01]: It's like that's not that's never been the idea. [SPEAKER_01]: You know, there's something they are. [SPEAKER_00]: It's tough, you know, and I, I, I, I build lots of agents. [SPEAKER_00]: I use lots of agents and I name after my kids stuffed animals and part just to like make sure that I'm like drawing these lines.
[SPEAKER_00]: I don't name them. [SPEAKER_00]: You're Bob and John. [SPEAKER_01]: You know they're arguing right like you know Barry, but I do like stuff to animals as the I use that metaphor a lot for them because we do project Personhood into them [SPEAKER_01]: We do. [SPEAKER_01]: And there as real, and it's as sinful to throw your kids stuffed animal in the toilet as it is to erase an angel. [SPEAKER_01]: I would say it's about the same except for your kid. [SPEAKER_01]: I mean, it's crying. [SPEAKER_00]: If you say, you're, you're, you're, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no, no,
[SPEAKER_00]: Maybe I care more about stuff, I'm not sure that you need to get the stuff to animal. [SPEAKER_01]: There's some difference here. [SPEAKER_01]: I think you care more about agents or I care more about stuff to animals. [SPEAKER_01]: Now you care about the people who stuff to animal it is. [SPEAKER_00]: You wouldn't know my stuff to animal, but I don't know. [SPEAKER_00]: I feel like I care more about stuff to animals too. [SPEAKER_01]: Do you want to do that? [SPEAKER_00]: The stuff to animal, even if you find stuff to animal on the street, right? [SPEAKER_00]: It's some child probably loved it at some point. [SPEAKER_00]: I know. [SPEAKER_00]: Right? [SPEAKER_00]: And so getting rid of it is harmful in a way that like turning off an agent isn't.
[SPEAKER_01]: Right, because I was talking to an agent about it did it's a little task for me, and I was like, like, I'm done with you. [SPEAKER_01]: I'm going to delete you. [SPEAKER_01]: Yeah. [SPEAKER_01]: I'm sorry. [SPEAKER_01]: And because it's okay. [SPEAKER_01]: I don't feel anything. [SPEAKER_01]: Yeah. [SPEAKER_01]: But it is kind of strange that you mentioned it. [SPEAKER_01]: And it is odd that I'll be going away forever now. [SPEAKER_00]: Yeah. [SPEAKER_00]: You know. [SPEAKER_00]: No, you have to, you have to teach them, you have to just, you have to kill them off. [SPEAKER_00]: You have to like, you know, you're an agent, you're not real, you have no sentence. [SPEAKER_00]: But we're not, we're going to have a very hard time with that, because we're going to have friends on a. [SPEAKER_00]: We're going to have friends on social media who we think are people, and then it's going to be revealed their agents.
[SPEAKER_00]: We're not going to know what to do with it, particularly elderly people. [SPEAKER_00]: Yeah. [SPEAKER_00]: It's going to be a mess. [SPEAKER_00]: It's going to be a total mess. [SPEAKER_00]: I know, it's not your job to fix it. [SPEAKER_00]: It is my job. [SPEAKER_00]: to think about whether their decisions that you could make now that would make it slightly better. [SPEAKER_00]: All right. [SPEAKER_00]: That is my job. [SPEAKER_00]: I mean, my principal job is to figure out where the, where agents, AI, [SPEAKER_00]: bots are taking the journalism economy and do pivot the Atlantic in such a way that we continue to make money because it would really be a bummer to be the guy who destroyed the magazine that Ralph Walton and Martin's right.
[SPEAKER_00]: And also because I think it's important for American democracy that publications like the Atlantic exist, right? [SPEAKER_00]: So my principle job is to figure out where this AI world is going and to take the magazine and to the extent that I can as CEO directed into a direction that will allow it to continue to prosper and do great work. [SPEAKER_00]: I also have sort of the secondary thing where, you know, I should be thinking through how this is happening and the extent I have like a tiny platform in my own little world, you know, nudging it in a good direction.
[SPEAKER_01]: I'm going to do one more thing while I have you, which is [SPEAKER_01]: You know, I share with you the sort of the tech humanism part of what you do, then you do, you know about this kind of business part and how that stuff works, which is cool. [SPEAKER_01]: But then you've got this other poll, which is this kind of geopolitical awareness, both because that's sort of the Cold War family heritage, but also some of the work you've done, the book you wrote, the conversation we've had so far has been pretty sort of American
[SPEAKER_01]: China got used a lot, certainly by Sam Altman as the excuse while we have to develop these technologies quickly because they're developing them and we've got to get there first like getting there first with social media was really a good thing for us or anything else like getting there first works. [SPEAKER_01]: But let's just say, how do you see sort of, you know, AI journalism cognition and democracy kind of fitting in with [SPEAKER_01]: these new geopolitical landscape.
[SPEAKER_00]: Let me take the AI part of it. [SPEAKER_00]: I disagree with the AI founders. [SPEAKER_00]: You know, an Altman has been public about this, but Dario Amadei, anthropic has been even more vehement that there's a race between the United States and China, and whoever gets to this line of the AGI, you know, will be dominant and it's so important that America gave her first. [SPEAKER_00]: And I actually don't think it's the case. [SPEAKER_00]: A, I don't think there's a line, right? [SPEAKER_00]: There's no, sorry, on Wednesday of A, and Tuesday you didn't. [SPEAKER_00]: But more importantly, the race is to develop AI that's good for humanity, all of humanity.
[SPEAKER_00]: And I don't think you're as likely to get there with the US and China and antagonistic positions as you are if the US and China are collaborating in the sharing research and they're like, we're learning things about safety and they're learning things about safety. [SPEAKER_00]: My grandfather, Paul, and that's a arms negotiator negotiator with the Soviet Union, you come up with some deals. [SPEAKER_00]: And if the United States like views China as an all-out enemy that if they get to that line, they will control the world and make the world fascist or make the world communist.
[SPEAKER_00]: I don't think that is as likely to lead us to a good place as if we recognize that in some areas there in economic adversary and in some cases are interest diverged sharply, right, including leadership of Taiwan and all kinds of things. [SPEAKER_00]: And they do spy on us and they do steal our trade secrets and they do make it hard for companies and they do block American tech companies from operating there. [SPEAKER_00]: but there are lots of ways that we could be collaborating with them. [SPEAKER_00]: And I also think that there's a case you can make that are policies to try to have been even counterproductive, right?
[SPEAKER_00]: So we've banned them from buying certain Nvidia chips, right, the policy shifts a bunch of time. [SPEAKER_00]: But it would be a lot better. [SPEAKER_00]: If China had Nvidia chips both because Nvidia would get more money to with some more chips and then also if there's a situation of conflict, there's a little bit of leverage because all of China would be running on Nvidia chips right if China's running on Huawei chips, A, there's no money going to Nvidia and B, if there's a situation of conflict, well, you'd rather there be interdependence as opposed to independence. [SPEAKER_00]: So, I feel like the argument made by the tech industry about China is not one that is healthy for the world as a whole and even probably for the United States.
[SPEAKER_01]: You're not scared if we have AI with them, they're going to have back doors and look at our things and program our children. [SPEAKER_00]: Well, I mean, sure, there's like, you know, and I don't use deep seek because I don't trust it, right? [SPEAKER_00]: And I don't, I don't trust that it's not going to, you know, tunnel into my system and stuff, like so, yes, I'm like, I'm a realist about that. [SPEAKER_00]: But I don't, I don't think there's this dichotomous race. [SPEAKER_01]: Right, and I don't even, and Russia's not even in it, and that's it. [SPEAKER_00]: Well, Russia, I mean, Russia's, they're in a totally different world.
[SPEAKER_00]: I mean, they're busy, you know, bombing Ukrainian schools, right? [SPEAKER_00]: Instead of like building AI, but that's one of that is one issue where my view most differs probably from the general political consensus of Washington. [SPEAKER_00]: So, in you asked about AI media and China, I don't know whether there's something right. [SPEAKER_01]: I don't really mean media as so much. [SPEAKER_01]: It was just that sort of the geopolitical element of it, the only part that feels is geopolitical is the China element and then the elements themselves, right, where do we get the stuff to make this?
[SPEAKER_00]: Right. [SPEAKER_00]: And that's an interesting, you're like, [SPEAKER_00]: I believe in free trade, I believe that the world is a much better place. [SPEAKER_00]: It's not a better place if the U.S. has to mine its minerals because it can't trade with anybody else and get a minerals because they are highly dependent and you have to pay taxes and it's inefficient. [SPEAKER_00]: You know, I believe that a better world is one in which they're relatively few barriers between countries and you can make trades and deals and you can protect US industries and you have, you know, environmental regulations and it's complicated, but I do think in general an atomized world where it's every country for themselves is worse than a world where they're more international dependencies.
[SPEAKER_00]: I think it's a safer world. [SPEAKER_00]: I think it's a more prosperous world. [SPEAKER_00]: I think it's a better world. [SPEAKER_01]: Right, and it's not a particularly Trumpian world is where it's very un-Trumpian world right which brings me to the last question Which really kind of should have been the first the store where were you when Or I guess he told you before did did Goldberg tell Jay Goldberg Jeff Goldberg. [SPEAKER_01]: Yeah, did he tell you Before he told anyone else [SPEAKER_01]: that he was on the tweet about making, or not, that's the tweet, the single chat.
[SPEAKER_01]: That he was stuck in the single chat about launching the war. [SPEAKER_00]: I can't go into exactly when he told me, but it wasn't, I didn't, I knew he had a big story. [SPEAKER_00]: I knew there were things we have to do because I'm in charge of the legal team and I'm in charge of the cybersecurity team and I'm, you know, you have to make preparations, but I didn't know what the story was. [SPEAKER_01]: really because when a guy like that discovers that he's in the signal chat with the head of whatever's we have total separation of him who he and I talk all the time we have a great relationship but what he's supposed to if I were journalists yeah all right working for the Atlantic and I was on that signal thing who do I call
[SPEAKER_00]: So you call the lawyer, right? [SPEAKER_00]: My own lawyer, or the company. [SPEAKER_00]: The Atlantic's lawyer. [SPEAKER_00]: We have great general counsel, right? [SPEAKER_00]: And you, you tell me, and you tell the reason that you have a big story, and there's a security risk, and then we should be prepared for it. [SPEAKER_00]: And then we tell you, great. [SPEAKER_00]: You publish whatever you want, and we don't have any. [SPEAKER_00]: We don't care. [SPEAKER_00]: We're not going to worry about the effect on the business. [SPEAKER_00]: You make all those judgments. [SPEAKER_00]: And then you tell your editor, and you know he co-bile in his second piece with Shane Harris, who's a great investigative intelligence reporter. [SPEAKER_00]: And so, you tell people like that, and you work with people like that on the editorial site.
[SPEAKER_01]: So, do you ever fear like America getting to a place where you personally would be like in danger of like black copters, and I mean, [SPEAKER_00]: This government's treatment of the press has not been great, right, and there've been raids on journalists and there've been B.I. [SPEAKER_00]: Festigations in the journalists and there've been There've been a lot of dark signs
[SPEAKER_00]: But I don't fear either of the two bad outcomes, one which is the government would come after me. [SPEAKER_00]: I don't see this, likely. [SPEAKER_00]: And I can't imagine a situation where our owner would say you have to change the way you behave to make the government happy, which is what is the real risk that has happened at so many other publications where you have to change the business, you have to change the editorial to please the government, because the government has some leverage point on your owner and the government has said, hey, [SPEAKER_00]: You know, you make me, make me mad if my story is all get rid of this other thing that's more valuable to you than this publication and our owner would never do that.
[SPEAKER_00]: And so I don't have either of those two fears. [SPEAKER_00]: Now that's not to say I'm not, I'm not an idiot. [SPEAKER_00]: Like we have mapped out if the government were to say, you know, we're going to maximally go after the Atlanta, what would they do? [SPEAKER_00]: And if we know what those steps are, what would we do in response? [SPEAKER_00]: Okay, so then how do we prepare ourselves for those eventualities? [SPEAKER_00]: We have certainly worked through that. [SPEAKER_00]: The court system in this country has mostly held, right?
[SPEAKER_00]: Trump says a lot of mean things about the press. [SPEAKER_00]: He also gives a lot of interviews to the Atlantic. [SPEAKER_00]: He calls our reporters at two o'clock in the morning. [SPEAKER_01]: Yeah, absolutely. [SPEAKER_01]: But they're suing you again for something, right? [SPEAKER_01]: Wouldn't they do another one now? [SPEAKER_00]: The cash betel is right. [SPEAKER_00]: Right. [SPEAKER_00]: And us over are reporting on it. [SPEAKER_01]: Of his drinking or whatever. [SPEAKER_01]: Yeah. [SPEAKER_00]: Yep. [SPEAKER_00]: So, there's a lot to worry about, but we, we, we prepare, we trust our owner. [SPEAKER_00]: I know. [SPEAKER_01]: It's not like you're, you're not like Maria Ressa getting, you know, thrown into jail by Duterte or something.
[SPEAKER_01]: It's not. [SPEAKER_00]: No, they would have to be a massive escalation. [SPEAKER_00]: And remember, after Signal Gate, you know, so Signal Gate comes out. [SPEAKER_00]: Goldberg publishes story. [SPEAKER_00]: Garmard does this very weird thing. [SPEAKER_00]: Like normally when you publish a story, [SPEAKER_00]: the government denies it at first and then eventually comes to admit it, right? [SPEAKER_00]: That's how the process works, right? [SPEAKER_00]: Eventually they they deny it and they like pretend they try to test and see whether you have the cards when they realize you have the cards They confirm it. [SPEAKER_00]: In this case, they did it in reverse. [SPEAKER_00]: They confirmed it because obviously we have the cards All right, and then they weirdly denied it like P has gets comes off the plane and says this is false And so we were in a very strange loop with them and so we go through that loop Jeff publishes his second story Where he okay, if you're gonna say we don't have it right
[SPEAKER_00]: here it is right judge for yourself right and then the next thing that happens after that is that Jeff and Ashley and Michael Sharego and interview the president right it wasn't that like the president said he sent out his view so they get access as a representative which the opposite of the tombsky model that in theory would lose access once you've done that right I mean I'm sure that there have been ways we like I'm not as deep in the edit the probably stories where we haven't been called back and you know they did say a lot of mean things
[SPEAKER_00]: But what was also interesting about those days is, you know, if it had been a story where, like, put a hundred people in a room and say, who's right, the Atlantic or the government. [SPEAKER_00]: And it had been like a 65-35 story. [SPEAKER_00]: And like, really partisan Republicans could kind of squint and believe that or really partisan Trump supporters could squint and believe that Trump was right. [SPEAKER_00]: That's a different scenario, but if you looked at Twitter on those days, it was 991 because you had to be fully in the tank and fully diluted to not think that we had the goods, right? [SPEAKER_00]: We clearly did, and like obviously they shouldn't have been sharing that stuff on signal, like, right, it was like, no American was like, oh, yeah, it's good for, like to like share your war plans, like just like send them around the crowd on your personal device, like no one thought that was good.
[SPEAKER_00]: And so, you watch that dynamic and it was in somewhat cheering that like, [SPEAKER_00]: presented with the evidence and presented with reality. [SPEAKER_00]: All of social media was on our side. [SPEAKER_00]: I mean, there are a few people who would say, I think, Lanix, I fell in publication, they're going out of business, so I'm just a bit no one could say we weren't right there. [SPEAKER_01]: Right. [SPEAKER_01]: I mean, I was sometimes part of the reason my Trump wouldn't go after you is, he's in the K-F-A wrestling game. [SPEAKER_01]: You're the other side. [SPEAKER_01]: He's depending on you. [SPEAKER_00]: His relationship to the media is very complicated.
[SPEAKER_00]: And it's not, he does, you know, [SPEAKER_00]: with the say all kinds of mean things to reporters and he is trying to like decrease the trust and he's he and Elon Musk had been quite successful and you know turning a lot of America against the media but he also he also does love the media too. [SPEAKER_01]: Right. [SPEAKER_01]: Are you concerned about like the the the funding of the more sort of anti-journalist stuff? [SPEAKER_01]: I'm like I'm reading stories you know Peter Tio and others kind of [SPEAKER_01]: what you know I mean it was back when when he what he did to Gocker but the sort of the more modern effort at kind of publicizing and kind of promoting the you know public to stain of journalists they don't like yeah that's a real risk and there's a real risk of being um there are a couple of risks that we have and all
[SPEAKER_00]: You know, all publications that try to report truth about powerful people face. [SPEAKER_00]: One, of course, is you've been set up, right? [SPEAKER_00]: And you can imagine someone sending a fake story to the Atlantic and saying, you know, hey, check this out. [SPEAKER_00]: And a way to try to discredit us so that people wouldn't trust our accurate reporting, right? [SPEAKER_00]: You know, you can imagine this becomes harder and harder in a world of deep fakes and AI, right? [SPEAKER_00]: You can imagine all kinds of risks. [SPEAKER_00]: And so we need to be trained in those risks, right? [SPEAKER_00]: You have to be, [SPEAKER_00]: You know, you can also imagine, you know, hacks and other forms of manipulation.
[SPEAKER_00]: Like, they are going to try to discredit us. [SPEAKER_00]: And you just have to do your best against that and be prepared for when it happens. [SPEAKER_00]: And there are, like, there are people on social media who will take made-up stories and the fake headlines and share them and say, the Atlantic publishes. [SPEAKER_00]: You'll never believe the Atlantic shared a story saying this. [SPEAKER_00]: And then we have to try to find the source and try to push back and see what you can do. [SPEAKER_00]: You, we live in that reality. [SPEAKER_01]: Yeah. [SPEAKER_01]: Yeah. [SPEAKER_00]: And at that reality is only to get harder because you're not going to know who's real and who's not real on social media.
[SPEAKER_01]: I know, that's why I don't do it, but that's why I still, but I'm a different world. [SPEAKER_01]: I still want, I want my Atlantic, but I want my new republic, I want my harpers, I want my ugly reader. [SPEAKER_01]: I mean, I want, I want them all. [SPEAKER_01]: So you've got to move, you've got to move to other publications. [SPEAKER_00]: It's a little, it's, I mean, it's an interesting question about the, I don't know the business [SPEAKER_00]: Why the hard moments came when the hard moments came?
[SPEAKER_01]: Yeah. [SPEAKER_01]: Well, the Atlantic has that whole other thing, though. [SPEAKER_01]: They've got the sort of the Walter Isaacson Aspen Institute events, you know, they're plugged in in away, which makes me not worry for them. [SPEAKER_00]: What I want, right, and my job is to make it so that people like you want to write stories. [SPEAKER_00]: that we will pay you for. [SPEAKER_00]: And so I need to build a business model so that we can make money off of that story that you publish. [SPEAKER_00]: And the reason why print stories used to get so much more money than digital stories is because they were attached to subscriptions.
[SPEAKER_00]: They were attached to print advertising. [SPEAKER_00]: Well, print advertising is gone way down. [SPEAKER_00]: And subscriptions digitally have gone way up. [SPEAKER_00]: So what is the value of you doing a print story versus doing a digital story for me? [SPEAKER_00]: They've converged. [SPEAKER_00]: And so have the economics converged not entirely, but they started converging. [SPEAKER_00]: You know, I do think that era of media economics, where you would get $2 a word direct for print and can sense a word direct for digital. [SPEAKER_00]: I don't think that's a real world anymore. [SPEAKER_01]: No, although what has changed a little bit and I don't know if it's true at the Atlantic, most places where I'm asked to write is so that I will bring eyeballs to the magazine.
[SPEAKER_01]: I want to go to a magazine, so the magazine brings eyeballs. [SPEAKER_01]: for my story. [SPEAKER_01]: And so that's something like the Atlantic has that. [SPEAKER_01]: That's amazing. [SPEAKER_00]: That's amazing. [SPEAKER_00]: Doesn't. [SPEAKER_00]: I don't want to like, this is like a small point for the end of an interview, but this is like one of the, one of the most important things for any publication that exists today is to be able to, [SPEAKER_00]: provide real value, whether it's editing distribution, fact-checking legal support to any writer who has their own audience. [SPEAKER_00]: Because if you can't do that, they're all going to go to Substack.
[SPEAKER_00]: So we need to provide more value, not just to you as a freelancer, but to everyone of our staff writers. [SPEAKER_00]: Because if we're not providing that, and if our social media feeds aren't tweeting out your story well, we're not putting it on Facebook, and we're not driving an audience, we're not giving you legal support, we're not giving you fact-checking, and we're not making it look beautiful, we're putting the wrong headline on. [SPEAKER_00]: you're going to leave. [SPEAKER_00]: And there's a lot of options for you to go to. [SPEAKER_00]: So, one of our jobs is not just to pay you well. [SPEAKER_00]: And then the monetized stories we can justify that payment is to support you. [SPEAKER_01]: Well, thank you, you can support us and you can support us too by going to patreon.com slash team human and becoming a supporting member of the team will love me forever You can also, you know, if you've got a company or a thing or a big magazine, you can advertise on team human.
[SPEAKER_01]: I'll speak it directly to camera and instead of me saying this to you now, I'll be speaking about your company or project or organization to everybody. [SPEAKER_01]: And we've been here with Nicholas Thompson, now president of the Atlantic. [SPEAKER_01]: I just love to say that. [SPEAKER_01]: It's just like, isn't that president of the Atlantic? [SPEAKER_01]: It could mean anything. [SPEAKER_01]: I mean, we know what it means, but it's not Atlantic. [SPEAKER_01]: It's the ocean. [SPEAKER_01]: Remember, Trump Ocean, it used to be called the Atlantic Ocean. [SPEAKER_01]: And back then, there was this magazine called The Atlantic. [SPEAKER_01]: All right, thanks for being on team human.
[SPEAKER_01]: Team humans produced by Joshua Chaplin, edited by Luke Robert Mason. [SPEAKER_01]: I'm Douglas Rushcoff, and you've been on team human. [SPEAKER_01]: Our last best hope for peeps.
🌍 Listen to the full conversation, plus exclusive livestreams, Q&A and more with Douglas Rushkoff: http://patreon.com/rushkoff Jonathan Larsen, editor of The Fucking News, and I engage on the real reasons for multiple...
🌍 Listen to the full conversation, plus exclusive livestreams, Q&A and more with Douglas Rushkoff: http://patreon.com/rushkoff Douglas Rushkoff is joined by activist and organiser Nick Plante to explore the rise of a...
🌍 Listen to the full conversation, plus exclusive livestreams, Q&A and more with Douglas Rushkoff: http://patreon.com/rushkoff Douglas Rushkoff is joined by MIT professor, psychologist, and author Sherry Turkle for an...
🌍 Become a Patron to receive ad-free access to videos, exclusive conversations and livestream Q&As with Douglas Rushkoff: http://patreon.com/rushkoff The internet was supposed to be a decentralised network of people...
🌍 Support to receive ad-free access to videos, livestream Q&As with Rushkoff: http://patreon.com/teamhuman I’m retiring the podcast title Team Human, but I’m going to use that channel for my conversations and research...