Machines talking to machines: The future of the internet

27:28We can't find customers anymore. We can't find readers anymore. All of our old stuff isn't really helping. And then also, you know, a bunch of marketers being like, oh, we have a chance to coin a term. Like, let's start with GEO or let's let's see what sticks. But yeah, I don't know. You have a lot of people who are experiencing this massive shift through the lens of like a traffic dashboard, which is, you know, again, like a weird way to experience the world through just like an analytics screen. But they're all watching the numbers go the wrong way. And they're looking for something that might help.

28:00And there's luckily, or depending on how you look at it, unfortunately, a giant industry of people who have been doing that job for a very long time who are... They don't want to lose their jobs. Pivoting. Yeah. Yeah. So it's basically just a new version of SEO that is based on optimizing articles to be picked up by chatbots instead of Google search. That's the definition. Yeah, definitely. For anyone who hasn't worked around a business or any sort of enterprise that has to worry about SEO for decades now, you have had to, at the very bare minimum, make sure that your website is like showing up on Google.

28:39So you have all these best practices and like how the page is built and stuff like that. It just makes sure that when Google comes to crawl it, it gets what it needs and it can show up when people search for things. When people talk about SEO, though, they're mostly talking about the version of that that's like, we need to, you know, become more visible than our competitors. We need to do these tricks and like write things a certain way or publish things a certain way so that Google gives us more attention. And every like Fortune 500 company has full time people who do this. There's like a$70 billion industry of people who sort of like mingle this with different marketing and help people get seen on Google.

29:16Google itself has a bunch of guidelines and they shift all the time. The search algorithms have shifted for years. They make all these announcements about like, oh, we're deprioritizing this kind of stuff or we're trying to get rid of this kind of spam. And they share as little as they possibly can so that people don't like try to game the system. Meanwhile, all these consultants and just people, you know, running their businesses or whatever are like, OK, well, I've heard that if you say all the keywords in this part of the page that it'll get picked up better. Or it turns out that they're prioritizing content with a lot of expert sources.

29:50So you need to make sure that all the text says that you're quoting an expert or something like that. A million different tricks like that. A lot of folk wisdom, a lot of made up stuff. A lot of stuff is kind of obvious and repackaged in a million different ways. but something that everyone who like works on the internet has to sort of think about including like everyone in media your your level of exposure varies but when i was running a website called the all for a while it was much more front of mind because you really have to think about the whole product and you're like oh like how does a machine see this thing that we're putting out there and in a lot of places that filters down to like changing the style of the work or the types of things that get published lots of very high quality news organizations have people sort of like, you know, shouting at various reporters to be like, hey, there's a lot of Google interest in this thing.

30:35Do we have anything on this? Which is, you know, for all the cynical assumptions that people have about media, maybe that one's slightly underrepresented in the, like, source of our problems. What GEO does is attempt to carry that logic over to a world of chatbots. And in the most important ways, it doesn't really translate at all. Because with Google, you're like, all right, here's this thing. And it's deciding how authoritative and visible our websites and posts and whatever should be. And we have some idea of how to like nudge that a little bit, but the best we can do is make sure that we're ready to be crawled and we're trying to get traffic.

31:15With most chatbot interactions, you're not really leaving the chatbot. So you can try to write your stuff in a way that gets absorbed and summarized and synthesized by chatbots more often than other things, which is kind of the core of this GEO concept. But the end result of that is still, it's kind of nothing. Like you're not doing it for much or you're doing it for much less traffic. These chatbots will, a lot of them have like search functions built in now. And if you use ChatGPT, you will see more citations and links than you used to. But again, like these are machines that summarize in real time.

31:50So even if they do find your website and say like, oh yeah, well, we found this great source for like a product recommendation, or we found this analysis of this story that you were interested in, or we found like the one incredible piece of original reporting on this obscure topic that you're looking for. The next step is we found it and we will summarize it for you in like a very concise and readable way and leave a footnote if for some reason you want to click through, which as anyone who's worked in media for a long time knows, that's just like a 95 % drop off at least. So you're competing for visibility for a machine that gives nothing back.

32:26Now, the long term idea here is that like, maybe the links become more prominent, maybe these companies become more nervous about liability. Maybe there are a lot of queries that they don't want to just answer. Maybe they also want to build big advertising businesses of their own, which maybe that works with, you know, sending people to outside websites or whatever. But it's also just like very obviously not the same thing to anyone who's used these products. Like one is a chatbot. You stay with the chatbot. You spend your time there. You always return to the chatbot. Google, like by the end of your time using your web browser, you've left Google.

33:00You've gone somewhere else. And with the chatbot, no, you are probably still in chat GPT and then you're putting down your phone. Like it's just a whole different thing. However, everyone who's like in this online economy is worried about this. This is where their attention is drawn because someone is offering something. Yeah, I sort of wonder, like, is this going to be as big of an industry as SEO or is it just like a bandaid solution that is not really going to work to delay the inevitable, which is just like no traffic for anybody or only direct newsletter? I don't know what it would be. Right.

33:34Like, is it going to really ultimately like serve these same people if we just keep optimizing in response to the changes when, like you said, they're giving us nothing? Right. I mean, one thing that came up, and this might be like sort of a, you know, hopium situation, but some of the consultants were saying like, all right, if your business relies on just getting a big audience, like maybe a news site, this is just a nightmare transition for you. And there's really not a lot you can do. Like there's no benefit really to optimizing all your pages so that a chatbot picks them up and repeats them in summary.

34:06if you are like a store or a brand and you sell things people do eventually need to like leave the chatbot or buy the thing and so a lot of the focus has shifted to that kind of stuff where it's like okay people are going to ask chat gpt or gemini or whatever like what rechargeable battery to buy and it's going to offer them a list or something and if you can get in that list then there's real that is more similar to how things work now than the like version where you're you're trying to get eyeballs and like, you know, get people to come and hang out on your website or something like that. So it's going to shift from like people trying to game the system so that they can end up in chatbot results to when inevitably the AI companies start monetizing through advertising through everyone paying.

34:55Right. Or I don't know if that's what you mean, but it's going to it's going to result in people paying to show up in these results probably. Right. Well, yeah, they've just sort of taken a step out of various like commercial processes on on the internet like instead of going to google searching for some product recommendations going to another website reading that website and then maybe clicking through to something through an affiliate link or whatever and ending up on amazon like this long chain where each step has like some sort of value to it but also a lot of waste and monetization and whatever the chatbot version is like okay we'll just like we will do that in the background basically and then just give you the end recommendation and then you just buy that which you know again this is one of these things where you you kind of can't argue with the the fact that that is streamlined but it is also it does again turn something that was previously a fairly active process into a pretty passive one you're sort of instead of going to the store and picking out your food you're opening your fresh direct app and being like reorder cheese sticks you know like it's a different like interface with the world it's a different way to like interact with with information but yeah i mean the geo stuff is worth sort of lingering on a little bit as a practice though because seo i was talking about it before is this like thing in the background that you know people who worked online know about this insider phenomenon but it's everywhere like the evidence of it is everywhere if you open a recipe website and you're like why are there 20 paragraphs here that use these same words over and over again before we get to the recipe.

36:21That's the product of SEO. You open like any news site and you're like, why does this news site have an extensive and like well-funded e-commerce operation? Like New York Magazine has the strategist, which is great for, but it's like, why does this exist in the first place? It's because you're capturing search traffic to then like monetize on Amazon. Why are headlines written the way they're written? Why are news stories, why do they read differently now than they did in a newspaper? And it's not just that there's more space, it's like, okay, this has to talk to a machine before it talks to a person.

36:52So it completely reshaped not just like the economy of the web, but just the texture of everything around you. It's sort of made the world feel more uncanny and less human online in all these subtle ways that we just get used to, but that are absolutely the product of like Google's design and like Google's incentives. And you end up with really absurd versions of this that, I mean, that are now like ancient history. It was the late 2000s when a company called Demand Media launched and eventually went public on the entire business plan of monitoring Google search terms and then just manifesting articles to meet them.

37:32You would sit there with a dashboard and it'd be like a bunch of people are searching for like sundresses right now. And so we are going to do a best sundress thing in the next 15 minutes, like with contract labor, like really dystopian stuff. But like, that was almost 20 years ago. Now everything kind of works like that. Everyone's doing some version of that. And that's just been incorporated into our ambient experiences. So SEO is this underrated factor in making the world feel just different to be gentle about it. And I do think that attempts to reconfigure around chatbots and generative AI and stuff, you'll see some versions of this.

38:08Like a lot of the consultants I talked to were saying like, well, we're getting some good results from lists. You know, if we make lists, it's easy for the chatbot to then sort of take the list and like reformat the list. And it interprets it as a ranking and authoritative and that'll maybe work. Or it really helps us to get quoted in highly ranked news organizations. So we really need to hustle and make sure that our clients are getting quoted or that people are getting quoted or mentioned. Because if enough sites mention us, then the chatbot will just assume that we're important and we will then have biased the chatbot in our favor, which is like sounds a little hopeless, but it's also kind of how things it is kind of how things work.

38:49It's not a bad idea, I guess, in the narrow scope of this conversation, but it's sort of a dark way of thinking about how all this should work and also just not something that's going to be easy for almost any other clients to do. Right. I was going to say, even if the chatbots decide to continue pulling from are like giving more weight to established verified media organizations and the giant sea of disinformation that's getting spewed out there, you know? Yeah. And it's really funny because it's like at any time in history, if you were advising someone about how to get more attention, you'd be like, get mentioned by trusted sources or whatever.

39:24And to have this then filtered down through the logic of SEO and GEO and whatever, to then have people being like, it's great to get mentioned by these prominent organizations because we believe they are still being scraped, although we're not sure. For now, they're still being scraped. And then that will translate into more visibility in this thing that, like, strips them of sources and strips them of context and whatever. It's like it is just the dregs of genuine, you know, useful attention. But that's kind of where, I don't know, that's kind of where everyone feels like they are right now. Like, even huge companies that weren't so worried about this because, you know, everyone knows who, like, the big life insurance companies are or whatever.

40:02Like they're not worried about showing up on Google. They're freaking out about this because they're just like, well, we used to have a few million people looking at our stuff and now we don't. And they're talking with chatbots that we don't feel like we have any connection to. What should we do? And what they're doing is like hiring people to analyze this stuff, to think about reformatting all their pages that no one looks at. And then they're also in response, I guess inevitably, using AI to generate just like massive amounts of what they believe to be AI friendly stuff. So you've got machines meeting machines, which is like a really common phenomenon in the world right now, but is happening at that scale too.

40:44Like companies that are just trying to be seen in the world so that they can do business are like, okay, we need to use AI so that the AI notices us. We need to use AI to publish like 7 ,000 pages that all look like FAQ guides or that all have diagrams in them. And we're just going to get that out there so that maybe the chatbot will notice us and mention us occasionally. I mean, not to like fall into the trap of personifying the chatbots, you know, in a new way. But the idea of all of these massive firms asking people for advice for how to just get a look from chat GPT is wild. But this is what's happening.

41:21I'm really curious what kind of journalism the GEO era will produce. Maybe it will just be sort of a flash in the pan and publications will just be like, oh, we're going to double down on some kind of like directly connecting with and serving our readers or whatever. But this is the SEO era produced a certain kind of journalism. It'll be interesting to see. Like, will hot takes come back? Like people offering controversial quotes or maybe literary quotes like that? That would maybe be not so bad. Yeah, one thing that comes up a lot is that these companies, they're all going wild right now and just giving people answers to everything, whether or not the chatbot can do it and whether or not the company that owns the chatbot really wants to be associated with those outputs or whatever.

42:05But there's like a whole bunch of stuff that's open ended. I mean, it's been interesting to me to watch this unfold. I think the timing was just right that this wasn't a massive deal during the last presidential election. but I think, you know, by now enough people use chatbots that it would be. But like you can ask any of these chatbots about politics. You can ask them what's going on with the election. You can ask them what they think of Donald Trump or whatever. And they're for the most part kind of careful and they're like, well, on the one hand and on the other. But they make these affirmative claims all the time and they are just the voice of these companies.

42:39And I do think that there's like a real need on the part of these AI companies to hand some of that off to people, not because people like do it better or more authentic or whatever, but just for the purposes of like liability and reputational risk. Like if you have a bunch of users, if like a third of your users will flip out if you say that Donald Trump has ever done anything wrong, what you then do is sort of say like, well, these websites say that maybe he did something wrong. These websites say that actually he did the best thing ever. And then you just sort of like hand it off like that.

43:11So I do think there will be pockets there that people will try to serve. So you might be right that there's like a take ecosystem that could exist there. But that also sort of feels like dregs. There's just not much. But the journalism that the SEO era produced, I think directly was uniformly pretty bad. The journalism that the SEO era supported, a lot of it was good. And that's the awkward thing about this is like you do SEO to like make your business work and to like keep your publication afloat. And you write stuff and publish stuff to like serve an audience that's coming from Google and doesn't care that much about you or whatever.

43:45But with the like economics of an advertising based website that ends up supporting things that maybe you do on purpose and that you maybe care about and maybe your readers actually like seek out and like this strips away a lot of that opportunity. I think there is an extra traffic to go around if there's low hanging fruit to be addressed about something that happened last night at an award show. For the most part, the chatbot will just tell you. And it won't get that from nowhere. It'll get that from people who've published about it and it will summarize it and synthesize it and whatever, but it will take what it needs and give it to its customers.

44:22And, you know, that's that.

44:26If you like what you're hearing on this episode, there's a place where you can join the conversation. It's called Cujoplex, and it's a private online hangout zone on Discord where smart, cool weirdos just like you, you know, people who like talking about the evolving state of independent culture, can congregate, share links, and talk about the news of the day. We've had some great conversations lately about creative economies, media theory, AI, politics, the future of journalism, and even the records and TV shows we've been obsessing over this year. Some of our chats have even inspired episode ideas.

45:02You can get access to Cujoplex along with our monthly bonus episodes by signing up for a paid subscription over at our sub stack. That's theculturejournalist.substack.com. What might our experience of the web look like in a world where conversational AI becomes our main portal to the internet, whether we sort of want it to or not? It feels like we're being kind of nudged in that direction. How do the big AI developers envision the future of web browsing and like internet use in general? And how might actual user behavior differ from that vision? For example, as you've reported, like Reddit, you know, a lot of people are retreating to Reddit right now or getting their information from Reddit for better or for worse, which represents like a much earlier kind of model of what the internet could be.

45:53Yeah, I think Google still pays lip service to this idea of like interacting with the web and whatever else. But really what most of these companies imagine the world to be like soon is people interacting primarily with the models, not leaving the chat window or the conversation. Like they really do think that that's just the new interface and that going through a search engine, opening a browser, entering an address, navigating as a concept, that's just gone. Like I think they really genuinely believe that that era is already in the past. When they do talk about like a wider world, when they do talk about a world beyond their specific AI models, they are using like the sort of loaded language of agentic AI where the AI is doing things on your behalf.

46:40They always use the same example as like a plane ticket, which is such a funny. The idea of an AI booking a plane ticket for you is plausible and like people are working on it and whatever. But it's also one of the funniest things to target because it's insanely annoying and hard. I have a friend who always brings up this example where he's like, imagine standing over someone's shoulder and like telling them how to click through Expedia and like book your ticket. And like imagine trying to do that through a human being who's like good at this. That's a huge pain. But anyway, so they imagine a world where you have agents like moving around the world for you, interacting with other agents.

47:14And so they do end up describing something that is like the web, but it is machines talking to machines, which like on a very basic level, I guess the web is machines talking to machines so that people can communicate with each other. This is a version where they are all connected in this proprietary AI way. You are sort of recomposing something like the web where you build a thing and it has an interface with the other things. And it's very chatty to the user, but behind the scenes, it's just software or whatever. But that's just like a totally nascent whole new world scenario that they're gaming out.

47:47And they're thinking in terms of like, all right, we're building these interfaces. We're going to get hundreds of millions of people using our chatbots. And then we're going to hook those into a bunch of businesses and we're going to take a toll somewhere. That's like a really interesting plan for them. And like maybe that'll work. But it really does not involve like active human navigation and like travel through a shared space. So it really is not, they don't envision anything like the web at all. And this, you mentioned Reddit, which is like this fascinating story at the moment, because I do think there's a real desire, even with like low stakes material, to like at least witness other people talking, if not like actually talking to them.

48:30to just sort of say like, all right, well, I'm thinking about something. I need something. I'm looking for something. Maybe I'd like to read some conversations between other people who have thought about this rather than, you know, reading an article or more recently, rather than talking with a chat bot. And so that, I think, has been a huge boon to Reddit, which, I mean, when Reddit was new and growing, it felt like the thing that was going to centralize and destroy all these small communities and like, oh, look, this is like a shitty version of forums and whatever. But now, in contrast to everything else, it does feel like it's fairly like well-preserved space, which really says a lot more about the web around Reddit than it does about Reddit itself, which is the same sort of like massively problematic thing.

49:15It is, in fact, a centralized community of communities wrapped in an advertising business, public company, all this, all the stuff that was a problem is still a problem. But in comparison to everything else, it does feel like a place where people are actually attempting to talk to each other. So that's a draw for it. But also, as Google becomes, like, shittier and shittier, the AI answers are taking a lot of the attention. The links are such trash now that they are just, like, answering a huge proportion of search requests with links to Reddit. They're just like, well, you might as well, like, you're asking Google, but we're just going to send you to Reddit, which has been great for Reddit, which is what's public.

49:52I really noticed that. And I trust Reddit. I found myself trusting Reddit more now and other results. Because I'm like, these are real people talking about their real lives. They've tested the thing. When's the first time that you thought about Reddit as a place and a community? What was the first reputation that Reddit had to you? That's a great question. The first time I thought of it as a place? Well, I feel like Reddit has been a lot of different things to a lot of different people over the years. But I have so many versions of Reddit in my head from 15 years ago. And I'm curious when it sort of entered your consciousness and what it seemed like before it was this.

50:27It seemed like a some sort of nerd website. Yeah. But then it grew on me and I started to be like, oh, this is really useful for lots of different things. I think for me it happened in tandem with watching again sort of the erosion of journalism and I saw it ascend in terms of finding out about more niche stuff that like news and culture organizations no longer had budgets to talk about necessarily. Yeah. And I trust it because it like I'm thinking of like even just like music fandom communities kind of or like finding. Yeah. For me, I like use it to find like live bootlegs and stuff a lot. Like that's when it sort of first emerged for me.

51:05So I guess I'm part of that nerd, that nerd wave. But it's also the fact that it's managed to be this self-regulating community, I think, is is part of like, you know, like we all like implicitly know, like the things that wouldn't be tolerated on Reddit. and that that gives it a sort of sheen of protection and trust or people know how to disagree with each other on there and nobody wants to deal with just like bot bullying and stuff yeah well i was asking because before it became like genuinely useful to like a wide range of people for like a wide range of stuff like there was definitely a time when reddit was like a reporting subject for me and also a place that i visited a lot where the idea that that would be a good place to find live music bootlegs.

51:47It's just not, it just wasn't. Like it didn't have that depth and it didn't have enough people on it. Right, it was almost kind of maligned a little bit. It was just like another like message board and like a lot of people were trolling, were trolling on there. I think that's worth talking about. Yeah, and also like I associated with its first big political movement, which was like supporting Ron Paul in 2008 or whatever. And I associated with a bunch of reporting that we did in like 2010, 11, 12, about how it had become like a haven for revenge porn and some of these really strong like normative features that it had where you build a community and run it according to your rules and like whatever were sort of directed toward stuff like that.

52:32And it was just like mostly young men with like specific priorities and emerging politics and like whatever. It was sort of a marginal thing. And now it is just the only place where you can start a community and expect people to show up, which is a wild, it's a wild transformation. They also, in addition to having tens of thousands of volunteer moderators that put in like a huge amount of time, they are spending way more money and attention on like making sure that it's a relatively hospitable place. And they have a lot less patience for people who are just like, oh, we're exercising our right to free speech here.

53:05They're just like, no, you're not. Yeah. Wikipedia is a lot like that, too, which is also voluntary. You know, I think there's like something to be said for like self-regulating, like voluntarily moderated Internet communities. Totally. It's these are some of the last huge ones. And they're both at risk in different ways. Like Wikipedia is absolutely devoted to its mission of being this nonprofit thing. But it has a really hard time getting people to keep contributing and to spend all this time editing. And I do think it is incredibly directly competitive with the way that a lot of people use chatbots.

53:38because Wikipedia for years was just like top Google result for all kinds of factual searches. And now the way the like AI stuff in Google works, but also just the way any chatbot works is like you just ask the chatbot and it basically summarizes Wikipedia, which it also just, you know, ingested and trained on. But yeah, it's sort of the last one. Reddit is at risk in a different way because it is both growing really fast and bringing in new people who are interested in like building communities there. And one thing that a lot of moderators there told me for the story I worked on recently is that even though you have a bunch of casual visitors who don't really think of it as a community and who maybe aren't like the most helpful, who are sort of like, where am I?

54:18Just like, you know, give me my content. you do have a lot of people who through Google are finding really niche communities who never would have thought to go to Reddit in the first place and who end up on like some sort of incredibly obscure sub forum and they're finding each other all like 200 people who care about something. So that's all really exciting. But also they are a public company. They're doing very well. They have thousands of employees when they used to have like 75 when they were big, like in 2016 or 17, as a tiny company. There's real money at stake. There are shareholders. They now have that like a new sort of need to grow infinitely.

54:56And they are growing really quickly. But it's an environment where you do have to wonder and some moderators sort of voice this concern to me. Like, why am I working for this for free? Why am I putting my life into this? I'm getting something out of it because I'm building a community. And I feel like, you know, it gives me, again, you'll hear this from moderators without apology, like a sense of purpose and power. you know like i have authority in this space but also at some point you're just working for a company that is like google or a company that is like meta and you're just like well i'm spending 20 hours a week on this and i'm like losing sleep and people hate me do i need to do this anymore so they do have that risk i don't know how long you can be at once like this specimen of very old web volunteer run, volunteer moderated community, a fast growing commercial web property with a ton of attention from Google that's making everything, you know, grow really fast.

55:52And also to exist as like a feast for all these LLMs that are like building a simulation of using Reddit, using Reddit, like by ingesting Reddit. It's just a wild place to be. But I mean, it makes sense that we're all thinking about that more. So, John, you alluded to this a little bit earlier in the conversation, But there's a common refrain these days that the internet will soon just be chatbots fighting with other chatbots, that social media is over because it's just bots commenting to bots, or in the case of the modern workplace, agents working alongside other agents. And, you know, usually we lament these developments in the context of a sort of futurism run amok.

56:33But you did write a column that frames these developments as more of an outgrowth of existing systems or even our frustrations with them. Essentially, these are the symptoms and not the disease. Can you explain that thought a little more? One thing that the sudden availability of all these like free AI tools that people a lot of times use without really disclosing it has had, I wouldn't say like the benefit, but it has certainly had the like effect of identifying extremely broken systems and relationships that are like not really founded on sort of good faith interaction or anything like that, but are instead just like a product of the technological background or whatever.

57:15I guess the most obvious version of this is what's happened in schools where you had almost overnight millions of students just like asking ChatGPT to generate essays and turning in assignments that were entirely AI generated. And this causes a massive panic, of course, because it's like a huge practical problem for teachers and everyone's suddenly concerned like, oh, my God, are these students learning anything? What are they even doing there? But that's, you know, a good question before AI got here. Like if if students understand themselves to be in an arrangement where the entire goal is to complete the text and submit the text and get the text evaluated, then one of the things that AI is doing here is just evaluating how hollow that experience is and how sort of like broken the incentives of education are.

58:01Like if students and teachers and everyone understood the point of being in a classroom as enrichment or learning for its own sake, and this was like well understood and intuitive to everyone involved, then this wouldn't really be a problem. It would be like a strange thing to do to submit your homework this way. But instead, it seems like an obvious thing to do. And so you have this incredible sudden x-ray of education as it exists in the world. Now, that's not to say that like the availability of tools that do this is just something that should be like tolerated and dealt with by all these systems that are vulnerable.

58:35But this does tend to highlight issues like that. I mean, we all can see it in workplaces, too, where if you have a certain workplace culture of people emailing each other and then suddenly people are emailing 10 times more. One of the things that you have to do in response is perhaps use AI to interpret those emails, use AI to respond to those emails. And then it is once again revealed like, oh, this is not workplace communication. Like we are not working toward a common goal here. We are engaging in some sort of like mutual performance in order to get paid. I mean, some of this stuff is it's not really helpful to point it out in the moment.

59:08It's like, yeah, of course, your job is like largely bullshit. But the way that AI heightens this very quickly, I think, is useful for thinking about a lot of the world, but also useful for thinking about AI. Like, what is it doing here? It's sort of like allowing people to break flimsy terms of all these different social arrangements we have and work arrangements we have. And if it is, in fact, that easy to break the deal, and if all it takes is having a text generator handy, then maybe it was kind of a rotten deal in the first place. But yeah, I do. I mean, to your point that there's this default assumption that the future is AI talking to AI and all that stuff.

59:44It's sort of true. But the point at which you have AI talking to AI, the point where humans are fully out of the equation is also arguably, at least in a lot of situations, a point where the interaction loses value. There are certainly things that, you know, machines can do by working with one another. But in a lot of work. Like finance. Like maybe it makes sense for like traders on Wall Street. Yeah, but even in those cases, those machines that are talking to other machines that maybe have some degree of autonomy, they're working on behalf of, like, people and firms with very specific goals, like, to further those goals.

1:00:18And they're also working against each other in, like, a really antagonistic way, which is something that I think is probably good to think about now. when, and we can, you can see the outlines of this in these scenarios we're talking about, where like suddenly you have teachers and students engaging in this antagonistic like tech arms race. They are in the antagonistic relationship. It is not the technology that's in an antagonistic relationship. The teachers are like, okay, well, I really don't trust my students now. So I'm going to subject them to like AI assessment and algorithmic assessment.

1:00:49And the students are like, oh, well, you don't trust us, then really fuck you now. Like I'm just going to find the best model for getting away with cheating on my homework. And you end up with like a total relationship breakdown between the people there, where AI is like an escalating factor, but it reveals that like that arrangement was implicitly or now explicitly antagonistic, but that it was implicitly antagonistic in the first place. In the world of like, you know, high frequency trading and finance, of course it is. Your people, like everyone understands this from the start because you are different firms trying to like make a dollar by exploiting the same resources or people like you're fighting over this thing that's very obvious to everyone involved and no one's pretending that it's anything else.

1:01:33But if you are in an office and suddenly like there's a supply shock of just email or presentations or whatever, maybe it becomes clear that your particular office environment is in fact a bunch of people who are being pitted against one another. Maybe that becomes harder to ignore. Maybe the performances that you put so much effort into, pretending that you are not being like subjected to this fully antagonistic experience all day, maybe that falls apart and you are sort of faced with the reality that you are in a workplace for reasons that aren't, you know, being a, you know, corporate family or achieving the goals of the firm or whatever.

1:02:11So I do think that you have elements of this all over the economy and in a lot of social environments too. like the story of like AI slop on Facebook is great for a lot of reasons. It's like a very visual example of the feeling that a lot of people have now of like all this weird generated shit appearing all over the place. It's so grotesque. There's like a little bit of like an opportunity for snobbishness about it because who uses Facebook anymore? It's all these like old people tricking each other with insane like shrimp Jesus AI generations or whatever. But I do think more than that, it's useful to expose very rapidly how horrible and antagonistic Facebook is as a system and how isolating it actually is.

1:02:55Because in a world with like shared norms and values and like a sense of actual community on social media, submitting a bunch of like engagement bait stuff, whether it's AI generated or not, would be seen as like antisocial or like a violation of the group norms or whatever. But that's not what Facebook is. Like Facebook is very explicitly an attention marketplace down to just like the way that it causes people to interact with their own families and their own friends. Like this, these sorts of arrangements are explicitly about, if not pitting people against each other, putting people into an environment where attention becomes this goal unto itself, where there are metrics and quantification for like otherwise just normal social interactions, if you then suddenly give people tools to just like spam each other at a professional level, it really shows like, oh, this was a fucked place.

1:03:47Like this space was ruined long before AI got here. And so I do think there's a lot of that happening now. Maybe this manifests at the same time that you're worried about your job because of AI, but maybe also you're just noticing like, oh, this office is fucked. Like everyone, given this fresh opportunity to behave in a completely like false, disingenuous or manipulative way to pretend they've done work they haven't done, to pretend they have expertise they don't have, to pretend they are someone they're not. Given that opportunity, they felt the need to take it. Everyone felt the need to take it.

1:04:24That's a fucked workplace. And so there's a lot of that happening with AI. I wouldn't describe this as exactly a benefit, but it is an interesting feature of this moment that with the like looming fears of automation as its own phenomenon, like this independent thing, you can miss or you should be paying attention to the various instances where people are using these tools to sort of leave behind or abandon any sort of pretense of values that they had before. Right, which kind of circles back to like what I was saying about how the AI chatbots and search engines are eroding our critical thinking.

1:04:59It's the exact same thing. Yeah, totally. So Mike Isaac from The Times recently published a trend piece saying that Silicon Valley has entered its quote unquote hard tech era, by which he means basically an evolution away from this sort of like employee perks heavy consumer tech focus, maybe B2B SaaS, Democrat leaning culture of the 2010s towards this culture more based in AI, data centers, a more competitive atmosphere, perhaps more right leaning or at least libertarian. politics. Have you, in your reporting sense, a similar vibe shift while covering this world? How would you describe it and how would you explain it?

1:05:40Yeah, I think what Mike is zeroing in on there is real and also related to what we were just talking about. In the arrival of AI as this sort of like cycle and also in the return of Donald Trump as president, you have for the leadership of tech companies, this like invitation to just abandon a bunch of pretenses. One that's been eroding for a long time, but that arguably sort of peaked like in early COVID era was that working for tech companies is fundamentally different. You do not work for a bank. You do not work in just any industry. This is like really core to at least the modern Silicon Valley mythology that you are working on building the future and this requires of you certain things and you are part of the special elect, but also it grants you special privileges.

1:06:38You are safe. You are in a different part of the economy than everyone else. You will get safe passage to the future. Everyone else will be left behind. But, you know, knowing that we should also have humility and we should care about social issues. We should foster a positive workplace. If we are modeling the future. It should look a certain way. And we recognize that your values as a bunch of like highly educated people who live near urban centers are probably aligned with this or that. We're going to sort of cater to these things. We are going to be the perfect like Obama era business exceptions.

1:07:13And I think the last gasp of that really was during COVID where tons and tons of people already at these companies recognized that, no, in fact, they did work at normal companies. They were just very successful and profitable, and they probably made more money than most of their friends. And they had this strange relationship to their companies with stock options and all that stuff. But, like, they were not doing God's work. You know, by then, there had been 20 backlash cycles for each of these large tech firms. What really changed after the, like, you know, hiring spree during the zero interest rate period is first companies realized that they weren't going to live in that world forever, and they probably had overhired and stuff.

1:07:49And they were looking for a story to tell as they trimmed back. You also had a lot of inpatients from management with their own employees who took some of this signaling about what kinds of companies they are quite seriously. And they would vocally raise concerns about like, what are we doing? Who are we partnering with? There were plenty of instances where employees raised concerns about Microsoft or Google working with the IDF, for example, before that was as salient as an issue as it is today. But you had versions of that all over the place where people would be like, oh, we're building general purpose AI tools.

1:08:21well, we should think about bias. We should, you know, think about how these might be used to make employment decisions or whatever. And basically, a lot of the tech leadership became really impatient after Black Lives Matter protests. They felt like their employees were out of line and talking in a way that was beyond their station. And then you have broader national political backlash, this conservative swing against and in tech. And you have the example of Elon Musk saying, I am upset with Twitter. I'm going to buy it and I'm going to destroy it and remake it in my image. And I'm going to fire almost everyone.

1:08:55Like that was insanely intoxicating to a lot of CEOs who aren't even who you wouldn't even think are like dispositionally or politically that's similar to Elon Musk. But in that sense, they saw what he did and they were just like, yes, thank you. And so I do think that a lot of what Mike is talking about in this turn toward like this doesn't have to be a fun place to work. Like you can't bring your whole self to work. Your job is, in fact, not safe. We can lay you off whenever we want. All this messaging is sort of like management regaining or gaining confidence and taking revenge a little bit and being like, listen, you guys, we weren't lying to you, but we weren't that serious either.

1:09:30Like we are in charge. You do work for companies. You do work in finance. Like you do work in marketing. You just have a job. And as soon as we have a way to fire you and we think we do with these AI tools that we are building, we will fire you. We will find efficiencies in your job and we will eliminate it. Like, I do think that for all the texture that Mike captured in like how this actually feels in the Valley, which is a profound shift that I think you'll hear validated by almost anyone who works there. There is just this like incredibly straightforward explanation, which is just managers are like, we are in charge.

1:10:06We can do whatever we want. Like, fuck our employees. You know, like, who do they think they are? You know, we're just about to automate them. Well, why do they think they can tell us anything that isn't their work product? Why do they think we care? So I do think that that is sort of like a story that is both very of the moment and specific to Silicon Valley, United States 2025. But it's also something that allows us to situate this in like more of a historical sense. Like this is management feeling like they have leverage. This is management like snuffing out labor organization after a few years of like flickers of organizing.

1:10:42and this is management sensing that they might have an exit in the form of, you know, an entirely new economy built around AI. And if they don't, they have so much cash on hand that what's the harm in firing a bunch of people and giving it a shot anyway? Like they're all so entrenched and incumbent that they can afford to spend like tens of billions of dollars attempting to fire everyone in the long term. Like, I don't know. They're really feeling themselves, I guess, is my read of that story it's like behind that story are a bunch of ceos who we think of as like different personalities and as competitors who are fighting each other or whatever but who are united in their like identity and position in saying like we are gonna really fucking go for it now like we are really gonna go for it we are free and it's funny because some of these guys read is like very timid and careful speakers or whatever but that is the impulse that is animating a lot of this and if They don't feel it strongly themselves.

1:11:41They recognize the desire from shareholders to, you know, flex those muscles a little bit. So, yeah, I don't know. I think that Mike really captured the cultural half of that. Although the one piece I would push back against is this idea that everyone is, like, re-devoted to building real tech. Because the, like, core work of building these AI systems is, in fact, pretty, like, narrow and contained within these companies. They still make all their money doing like incredibly not hard tech stuff of just subtly optimizing for ad clicks and stuff like that. So I do think there's a risk of overstating that a little bit because that does it does sort of make it sound like they've all shifted to like a Manhattan project or whatever, which in terms of like spending and stuff they have.

1:12:25But they're just they're buying Nvidia chips. They're buying real estate and they're hiring like a few hundred people from a small talent pool to see if they can push through, which is a significant thing, but is not necessarily the same. It's like saying, hey, we are Google. You know what? We're going to go to the moon now and we're really going to mean it. But yeah, it's a moment. It's a viscerally felt in Silicon Valley and comes up to my reporting all the time. Maybe like the AI stuff or the hard tech, it's maybe a shift in what is culturally valorized, but it's not necessarily what the reality is.

1:12:54Yeah. And another thing that Mike brings up is that there is this openness now in Silicon Valley's pursuit of military contracts and stuff like that. And that was a big part of even the prospect of working for Elon Musk at like a company like SpaceX, for example. It's like, oh, you want to work on rockets? You want to work in aerospace, but you don't want to build weapons? Come work for this new kind of company that would never do something like that. And then it's like, all right, well, OK, you are as embedded with the government as any company could be for a time. You are like openly fantasizing about, you know, intervening in wars and in some cases directly like affecting the outcomes of armed conflicts.

1:13:34Like this is not really a meaningful distinction anymore. The idea of Google, this like late 90s search tool turning into a military contractor would have been absurd at the time. And then by the middle of the company's life, it would have been anathema to this culture that they like cultivated and sort of like conveyed to their employees that we are, again, like a different kind of company. You're in a different kind of industry. We're building this like soft and wonderful future. We're going to do all this cool stuff and we're going to have a conscience while we do it. It would have been crazy.

1:14:04And now they're just like, we are working with our allies on war. You know, like, what else would we do? What do you mean? What do you mean you're upset? Like, get out of here. And, you know, you now have people who were pushed out of the company years ago over voicing concerns about, like, how future AI tools will be used. They've been out of the company for, I mean, this has been true for a long time. But now I think the open embrace is pretty jarring for people. I think that's a real change. and, I don't know, healthy in the extremely long view that these companies are more honest about who they are and what they have to some extent always been, but nonetheless sort of distressing in the near term.

1:14:47Well, John, thanks so much for sharing your insights and for coming on. Where can people follow you and find your work? I am at New York Magazine. The column is called Screen Time, but no one knows that. It doesn't matter. It's such a good column name. I think it's probably like 10 columns called Screen Time. It's so good. Yeah. I'm on Blue Sky posting links. I'm really on, I don't know. Anyway, thanks for having me. Thank you. Thank you.

1:15:22This episode of The Culture Journalist was produced and edited by Emily Friedlander and me, Andrea Dominic. Our theme music is by Mark Donica. To check out more, head to our Substack. That's theculturejournalist.substack.com. And if you like what you're hearing, leave a rating or review on Apple Podcasts or share us with friends to help support independent journalism.