Offline with Jon Favreau - A Techno-Optimist’s Case for AI

Episode Date: December 6, 2025

Economist and techno-optimist Noah Smith, author of the Noahpinion Substack, joins Offline to debate the promise of artificial intelligence, the benefits of online fragmentation (could it be good for our society?) and whether liberal nationalism is feasible—and a good thing. Though Noah and Jon differ on a lot of “Offline” themes, they find common ground on the dangers of social media, leftist scolds, and a country with an identity crisis.For a closed-captioned version of this episode, click here. For a transcript of this episode, please email [e-mail] and include the name of the podcast.  Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Offline is brought you by Quince, Cold Morning's Holiday Plans. This is when you need your wardrobe to just work, and that's why I'm all about quince. They make it easy to look sharp, feel good, and find gifts that last. From Mongolian cashmere sweaters to Italian wool coats, quince pieces are crafted from premium materials and built a hold-up without the luxury markup. Quince makes the essentials every guy needs. In addition to those sweaters and wool coats, they also get some denim and chinos that fit just right. Their outerwear lineup is no joke, down jackets, wool top coats, and leather styles that are

built to last. Each piece is made from premium materials by trusted factories that meet rigorous standards for craftsmanship and ethical production. By cutting out middlemen in traditional markups, Quince delivers the same quality as luxury brands at a fraction of the price. It's everything you actually want to wear built to hold up season after season. Love quince. Go to their site all the time. Buy some clothes. Italian wool coat. I could use a coat, actually. And maybe some gift shopping. It's holiday season, so go check it out. Get your wardrobe sorted and your gift list handled with Quince. Don't wait. Go to quince.com slash offline for free shipping on your order and 365 day returns. Now available in Canada,

too. That's Q-U-I-N-C-E.com slash offline. Free shipping and 365-day returns, quince.com slash offline. At Radio Lab, we love nothing more than nerding out about science, neuroscience, chemistry. But we do also like to get into other kinds of stories, stories about policing or politics country music hockey sex of bugs regardless of whether we're looking at science or not science we bring a rigorous curiosity to get you the answers and hopefully make you see the world anew radio lab adventures on the edge of what we think we know wherever you get your podcast almost all of the discourse around AI is actually people realizing the problems with the internet and taking this out on AI just like all the sort of positive

about the internet, a lot of the positivity of the internet was people feeling good about how TV and radio and whatever turned out. I'm John Favreau, and you just heard from economist Noah Smith. Know as many things, a former finance professor and Bloomberg columnist and the author of the Noah Opinion Substack. But one reason I wanted to talk to him is he's a real techno-optimist, and I am well aware that's a specific viewpoint in short supply on this. his show. A lot of nose takes, in fact, are pretty different from my own, that online fragmentation might actually be good for our society, that short-form video is a good alternative to social media, and that the dawn of AI is not going to end in disaster. But his substack pieces are thoughtful, measured, and well-research, so I wanted to have him on to make his case, especially

on AI, where he's quite bullish, so that you can all hear another perspective that isn't so Dumerish. After our conversation, I'd say I'm still closer to the Dumerish end of the spectrum, but I thought he made some good points and gave me a lot to think about. We also talked about the ills of social media, why he thinks the era of social media is finally ending, and why that might be good for American democracy. Here's our conversation. Noah, welcome to offline. Hey, great to be here. So I'm a long-time reader of your substack. I've wanted to have you on for a while because you write about

quite a few offline related topics, AI, social media, the phones, the political moment we're living through. I want to touch on all of that, but I thought we could start with one of your most recent pieces titled, I love AI, why doesn't everyone? That's right. So I definitely find AI quite useful for research and other tasks. I am not a full AI doomer, but I just have almost no confidence that we are prepared to handle the potential dangers and disruptions that are coming.

But give me your best case as to why I'm wrong. I don't know that you are wrong. I think we're not prepared to handle those things. But I think that's the norm. Like if you look at, were we prepared to handle the dislocation of local communities that came from the adoption of the automobile? Were we prepared to handle the changes in warfare that resulted from aviation? Where we prepared to handle the pollution from industrial chemical manufacturing?

were we prepared to handle the social changes and disruptions from the internet, were we prepared to handle, you know, or radio or even print for that matter, like the printing press, were we prepared for the printing press? Like, what's a good example of a technology that we were pre-prepared for, that we, you know, where none of the stuff ended up being disruptive because we knew how to handle it. And our institutions just like, yeah, we got this, man. We like, everything is good from this.

Like, what's an example of that? Yeah, so that is true, for sure. I wonder if it is what's, different about AI is just the potential for, you know, I mean, even the people that are developing AI, you know, talk about it like a humanity transforming technology. And the scale, I don't know if we've had a technology with the scale of the potential job loss, the potential threats. I mean, they're talking about bio-weapons. They're talking about, you know. And so I wonder if not being prepared for this technology because it could be so transformative is a bigger issue

than in the past. And I don't know, there's part of me that just feels like we're like rerunning the social media age experiment with the technology that's much more powerful. And I'm not sure that we've learned any of our lessons yet. Yeah. The idea that societies learn lessons and then those lessons stick. I mean, once in a while, you'll see some things a little bit like that, but I don't think that's typically true. Like, I think that we learn lessons in, in terms of our institutions adapt to previous technology. So right now, like, you're not scared of TV. You're not scared that TV is destroying the minds of the youth. But when I was a kid, people were scared of that. And before that, people were even more scared of that. Yeah. You know, but right now,

we've, we've convinced ourselves that TV doesn't destroy the minds of the youth. I just think that that people who expect us to like, you know, be prepared for this. I want that to be articulated. Like, what does that look like? What would that look like? What would be different than what we've ever seen? Well, it's like right now the, you know, the Trump administration is trying to fight states being able to regulate AI on their own.

And I do think there's arguments for having like one standard federal set of regulations for a technology like this and not having like 50 different states doing it. but I have not seen any effort, serious political effort, to put in place the kind of regulations and rules we might need to make sure that we, for example, are ready to handle the massive job disruption that could come from this. I guess we should start there too. Like on the economic side of it, like where are you on sort of the level of potential job displacement we might be facing here. It's interesting because when you say job displacement, what do you mean by that? You know, like Emote, the CEO of Anthropic says something like half of all entry-level jobs

could be replaced by artificial intelligence. Elon's out there. I think he said that at some point humans won't be needed for any jobs. Bill Gates has said something similar. Let's take a previous example of job displacement. So if you go back to 1800, let's say, almost all humans, I would say that 90% of humans or something like that worked in one job, which is agriculture. Almost everyone is a farmer, right? Everyone worked on these farms. Then mechanized agriculture comes and now almost no human beings work on farms.

So in the sense of like, if you just look at jobs that got replaced by this technology directly, it was, the answer was almost 100% of all jobs got replaced. And this obviously caused disruptions because then you had, you know, the dark satanic mills and you had Marxism and you had labor conflicts and you had pollution. And obviously, working class conflict and, you know, with management and bosses. And you had unions, management fights and like Henry Ford getting gangsters to shoot union guys on a bridge, you know, and you even got communism, which killed millions of people, you know, resulting from sort of the disruptions caused by the shift.

out of agriculture. And so I won't say that, oh, we handled that well. Everything was fine. Everybody got a nice factory job and everyone just, you know, yay, you're not a farmer anymore. I won't say that it was free of disruption. I mean, like, you know, tens of millions of dead, you know, citizens of communist China and the Soviet Union, not to mention like everywhere else, would beg to differ. However, humans still did important things. Ultimately, wages went up. Humans could make a lot more money doing the stuff they later did. So you could say that humans as a thing weren't replaced, but then jobs were displaced. You know, people were displaced from jobs. And I'm asking the people who were worried about AI to be explicit. Are you talking about another case of that? Or are you

talking about humans becoming like horses and nobody needs you and you're just this obsolete thing? I mean, perhaps. It seems like some of the, some of the predictions that it's not just from like Doomers saying that. I mean, like I said, you got Elon and Bill Gates saying stuff like that. The need for human labor itself might be replaced by artificial intelligence. Is this the first technology where that's been like a big worry, big widespread worry? That humans themselves might be replaced? Yeah.

The human labor itself might become obsolete. I don't know. You tell me. I can actually tell you. The answer is no. Yes, yeah. In the Industrial Revolution, tons and tons of people. At every wave of industrialization that happened, there were tons and tons of people who said there's nothing left for humans to do.

You know, from the Luddites that smashed looms to like, you know, people in the age of like internal combustion electricity, it said, look, human power is being replaced by machines. Like humans have nothing to do. And then humans found other stuff to do. We've, you know, in fact, more humans worked in the market than ever before. If you look at the percentage of humans that now works in the market, you know, we've automated much of manufacturing. we've now automated all of agriculture but if you look at what percent of humans have jobs what we call in the market you work for a company i'm not talking about like washing your dishes at home right uh then it's it's near record highs everywhere you look so the sunny scenario

here is that there is a period of huge and painful disruption like there have been in in the past when there's a brand new technology but that even as powerful as this technology may be, humans will find ways to do productive things that the economy needs. That is one scenario. And I'm not saying that's necessarily the most sunny scenario. There's also the scenario that humans keep doing most of the same stuff they've done before and just your job at work changes a little different. Your tasks change a bit. So like, you know, the internet, the internet, like if you look at what people do at jobs now, now, it's different than what people did at jobs in like 1985, right?

It's different. You don't have like the outbox, inbox. You don't have like these blueprints you're drawing. You don't, you know, regional salespeople are kind of dead, right? Yeah. Like Arthur Miller, death of a salesman. That's all gone. But yet you still have a lot of sales jobs and you still have a lot of, you know,

people still make spreadsheets for stuff. It's just electronic spreadsheets. And like people, you still have a lot of similar kinds of jobs, right? The internet wasn't a giant job destroyer in terms of, of like people didn't have to switch like they did in the industrial revolution. Like everyone had to switch out agriculture. Right.

But then in the internet, in the information revolution, let's say, you didn't see a lot of that switching. So I think scenario one is like people just do somewhat different things at work. You use AI now at your job, but you still have pretty much the same job. Maybe you're more productive.

You have to take on some more roles because like you have more time now because AI saves you time, blah, blah, but your job is generally easier. Now your job is like punching stuff into AI, correcting AI things, managing the AI agents. That's possible. That's, that's possibility number one, I would say. I can't tell you that's not going to happen. Neither can't Dario, man. He didn't know. He doesn't. He doesn't know. He's just,

forging ahead. How could Dario know? So that's possibility number one. I think possibility number two is what you said. There's this big disruption like there was when agriculture gave way to like manufacturing and services, right, where people just found different stuff to do. But there was a lot of disruption and society was disrupted and took a long time. So that could happen to. Possibility number three, this is the end of human usefulness for most humans, and most humans are just, like, useless. There's a lot to think about about that scenario. There's a lot of subtleties that people who think about that scenario don't actually think about, and how that would manifest and why horses became obsolete or things like that. parties to festive toasts by the fire. The holidays are full of moments to share a drink with

friends and family. Whether you're celebrating a year's end win or just relaxing after a busy day, being ready for the next morning still matters. Zbiotics, pre-alcohol probiotic drink is the world's first genetically engineered probiotic. It was invented by PhD scientists to tackle rough mornings after drinking. Here's how it works. When you drink, alcohol gets converted into a toxic byproduct in the gut. It's a buildup of this byproduct, not dehydration, that's to blame for rough days after drinking. Pre-alcohol produces an enzyme to break this byproduct down. Just remember to make pre-alcohol your first drink of the night, drink responsibly, and you'll feel your best tomorrow.

I love Zbiotics. I just bought 50 of them to try to get me through the holiday season and all the people who are visiting us. That's smart. I got my subscription, so it just arrives all the time. I need to double up on that. Make the most of every toast this holiday season. Just don't forget to bring pre-alcohol along for the ride.

Go to Zbiotics.com slash offline to learn more and get 15% off of your first order when you use offline at checkout. Zbiotics is backed with 100% money-back guarantee, so if you're unsatisfied for any reason, they'll refund your money no questions asked remember to head to zbiotics.com slash offline and use the code offline at checkout for 15% off off what do you think some of the subtleties people aren't thinking of in that scenario are one subtlety is like suppose AI produces all the stuff okay you just like AI agents are like producing all the houses all the cars all the dur-da-d-dur you know the software and the the, you know, health care and the legal stuff and who's consuming it? Is it just like five guys who own the AI companies? It's just like Sam Altman, Darya Amadei, Elon Musk. Is it just like

those guys using the power of AI to like colonize Mars and like create robot cities filled with nobody? Like just filled with robots, filled with like sex bots or something? Like what, like who is consuming all this stuff? Who's buying it? And how are they buying it with what money are they buying it? You're people envisioning this, this world of incredible AI generated abundance where AI just makes things much more cheaply and easily than humans ever could, makes all these different things from services to goods, everything in our world. You know, who's buying it?

That's what I want people to think about. Yeah. Consumers with their stipend that they get from the government for their universal basic income. Maybe so. Yeah, maybe so. Maybe we do that. But suppose we don't do that. Suppose we decide not to do universal basic income.

We just let the economy run. No, like, government check. And nobody has a job, you know, and so AI can make all this stuff incredibly cheaply. Who will buy it then? Why would AI make all the cheap stuff? Right. And so then your contention is, at some point,

there's going to have to be a market created or else. Like, are we envisioning a world where just AI doesn't create anything and everyone's super impoverished because nobody can buy anything so nothing gets made. But the minute a human tries to make something, AI swoops in and makes it instead, and then the human can't buy it. And then we're all just like, there's this equilibrium where, like, nothing gets created at all because AI could create it so easily we might as well not. And we all die of starvation. That's goofy. That's goofy.

That's not going to happen. So I guess the other thing that could be different about this transformation is the speed at which it happens. And so to the extent that you do have job displacement, it happens much quicker than it has in the past. And then you just have suddenly over a period of several years, I guess, a decade where you have a whole bunch of people who just like can't find work. And then I don't know. Is that something that concerns you that the fact that it's that it's going to, that the transformation could happen quicker than in the past? It certainly does concern me. It concerns me because a lot of the disruptions we had from the age of agriculture with people moving off the farms, moving into the cities.

like we didn't necessarily have a smooth system for those people to immediately all get good jobs in the cities. A lot of times you had slums, you had periods of high unemployment, you had overcrowding, you had a lot of things. And you had a lot of anger among the working class that boiled over into communism sometimes. I've seen something a little bit like that that happened 150 years ago, 100 years ago. We've seen that happen, but it wouldn't necessarily happen the same way this time. So I can't exactly imagine, I've read science fiction stories about like all the, angry truck drivers start rioting, you know, because of the self-driving trucks. I can't tell you that won't happen, but then, like, that seems like a pretty thin,

limited, like, could you really imagine something like Mao Zedong from 1850? Like, even Marx didn't imagine all the crazy shit that would be done in his own name. Like, he didn't imagine any of it. He talked about dictatorship of the proletariat. He didn't imagine what Stalin would look like. And so you couldn't imagine. You know, you could be dystopic and, like, I could tell you that, oh, there'll be like these robot lords that will, like, surveil us all with our armies of robots.

and we'll just be cattle to them. I can spin you of them wildest sci-fi scenarios, but I can't tell you the probability that any of these things will happen. Yeah. Like I said, I don't quite know either. I do feel like there's, you know, we already have a lot of that sort of populist anger brewing

in our country, in many countries, in different political systems. So I do, I sort of wonder what happens when you sort of throw fuel on that fire with the AI transformation, you know, that's coming and if that makes things a little worse and again like I don't I don't know what the solution is because I don't think you can just like stand in front of the technology and say stop right like it certainly seems like we're not going to we're not going to slow it down anytime

soon but you know then you hear people like you know Tucker Carlson people like that be like we should just ban the self-driving cars for that reason for the truckers I mean Europe might do all these things but then that's just going to make them a backwater Europe like relies in the car industry they're going to export zero cars and they're going to only be selling cars to each other because they mandate that you have to have human-driven cars, so they won't, don't allow self-driving cars. So they just, Europe can only make and sell cars to Europe, but they can't sell cars overseas, which means they can't get foreign exchange to buy the materials to build their own cars for

themselves, build their own human-driven cars themselves. They can't get the foreign exchange, you know, because they're not selling cars overseas. And so they can't get that foreign exchange. And so what happens then is that their costs go up. And European manufacturing basically vanishes. You're already seeing a lot of this with electric cars out competing, like the German car industry. Today, with self-driving cars, it would be 10 times more. And so, like, Europe just becomes, they become large North Korea because they aren't able to trade to get the stuff they don't naturally have.

Because they can't sell anything the rest of the world wants because they've banned it all. And so, like, we could become large North Korea. And perhaps there is something fundamentally humanist in keeping ourselves prisoned in the worst regimes of the 20th century forever, rather than simply embrace the uncertainty of new technology. What is your ideal public policy to handle the potential job disruption? Because I think we've all seen that job training programs, I think they have a spotty record over the last several decades. The most effective policy that we know of to help people switch jobs is not training

assistance, but job finding assistance. So training usually happens at the corporate level, or sometimes you go back to school. But, like, training happens at the corporate level. You just learn on the job how to do stuff. That's how most corporate education happens. Denmark has this system called Flex Security where basically, like, what, you lost your job? We'll find you a job. They have, like, basically, government-sponsored recruiters.

And Japan is trying to copy this. And so that could be an effective system for helping people find new jobs. That doesn't mean that you don't have to have government bureaucrats who know exactly what kind of thing you'll be qualified for, you know, or how you become qualified for that. But then they can, you know, if you're just sitting there, like thinking, should I send out resumes and stuff? It can be very demoralizing. You can sit on your butt for a long time. Denmark is like, we'll help you, you know, we'll take on some of this burden for you, we'll assist you. That doesn't mean they know what you'll eventually find and they'll know what you can eventually do.

They're not like perfect omniscient matching service. Right. But they can assist you with job search. They can provide that assistance. And that can be very valuable. So that kind of thing, I would love to see that kind of thing in the United States. How do you think the education system here would need to? change to help that?

I don't know, because I think there's so much of the education system that's going to change anyway, I think AI tutoring is going to be the first technology we've had that can take education out of the hands of, like, large classes. So past a certain size, like class size doesn't really matter. Essentially, one-on-one tutoring is much more effective than anything else. And once you get to, like, just three or four people, like, you might as well have 50 people. Like, class size doesn't matter. So we've got technologies that can increase or decrease class size, whatever, but we

We can't. So far, we haven't any technology that can scale up tutoring, private one-on-one tutoring, which we know is by far the most effective education method. It's just not economically efficient because who can afford to hire a tutor, only rich people. Like, I've been a tutor. I know it works, but I know it has to be one-on-one. And all the research bears that out. So instead, what we may be able to do is we may be able to give people AI tutors. We're not quite there yet, but we may be able to do that. There's no reason in principle why that shouldn't be possible. And we give people ad tutors and then people just learn much more. So that may be a quantum leap in education. In terms of what things people should get educated about, it may be that AI tutors are able to

flexibly figure that out by doing searches. AI can search the web and process an unimaginable amount of information and kind of analyze it. So maybe that AIs are sitting there saying, well, here's what you should learn. This would be a good idea to learn. And that AI itself will solve this problem or not solve it, but at least address it more effectively than like human created institutions could do. So you may have an AI tutor a little like, imagine you have like a little like primer, illustrated primer, you know, it's AI and it's an AI agent. And it just, it can educate you, but it can also figure out what you need to learn. And potentially what you need to learn and where you could be useful and what kind of job that you could get in this. That's right. And it could do it

a much higher frequency than like the school system. You know, imagine if your, if your tutor were also like a personal agent, labor market agent for you that could just like find you stuff to do for work. Wouldn't that be fucking cool? That would be cool. Yeah, I'd like that. This is more of a short-term thing. How worried are you about an AI bubble and one that potentially leads to a financial crisis?

I know you wrote about this recently. I'm not actually worried about the implications of that for AI itself. I think that if you look at history, you see this happening again and again with these new technology buildouts, telecoms, railroads, you know all the standard examples, right? In all of those cases, there was a giant bust, but the bus didn't stop the technology itself. Like all those railroads that got built got used. All those fiber optics that got laid got used. And so I don't think we're building lots of useless stuff that will stand like an old Soviet monument out in the tundra.

You know, like data centers are not going to be like these useless monuments. That stuff will all get used. AI works, you know, people keep investing in it. Like, the day that crash happens, I'm buying all the AI stocks. I'm ready to make a little money. Ready to get a free doubling of all my capital. Offline is brought you by Delete Me. Delete Me makes it easy, quick, and safe to remove your personal data online at a time when

surveillance and data breaches are common enough to make everyone vulnerable. Data brokers make a profit off your data. Your data is a commodity. Anyone on the web can buy your private details. This can lead to identity theft, fishing attempts, and harassment. But now you can protect your privacy with Delete Me. Your name, contact info, social security number, home address. Even information about your family members could all be compiled by data brokers and sold online.

The New York Times Wirecutter has named Delete Me, their top pick for data removal services. Yeah, so Delete Me is good to, you know, we all have way too much information out there online. I know I do. And so it's really important to just have a service like this because you don't want to go through all of it yourself. you use DeleteMe, takes it all off for you. It's very easy to use and very helpful and thorough. Take control of your data and keep your private life private by signing up for Delete Me. Now at a special discount for our listeners. Get 20% off your DeleteMe plan when you go to JoinDeletMe.com slash offline and use promo code offline at checkout. The only way to get 20% off is to go to Joindeletme.com slash offline.

Enter code offline at checkout. That's JoinDeletMe.com slash offline, code offline. So Anthropic stress tested all the big LLMs earlier this year, found that it wasn't too difficult to give them prompts that led them to attempt blackmail, espionage, even homicide. What do you think about like the potential for either self-regulation for some of these companies to try to figure this out or like government regulation? And I realize this is extra challenging because say we had a fairly robust regulatory regime here in the United States. You've got China doing their thing. You've got individual loan actors doing their thing potentially. Like another big concern is just sort of the,

you know, the robots will kill us all, uh, concern, which is a, yeah, there's a lot of concerns. Yeah. So, so every technology gets used by bad actors. Right. The USSR got nukes. They didn't use them, but they certainly used them to blackmail people into not stopping some of the more nefarious things they were doing. You know, bad people got a hold of industrial technology and used it to like wipe out whole cities. Yeah. Like, uh, it's hard. to name technologies that didn't get used by baddies for bad things. Yeah. Even agriculture fueled the rise of like large armies.

People will use technologies, new technologies, they'll use it to like whale on each other. They'll use it to beat the fuck out of each other. Yeah, but like military stuff, terrorism, oppression, surveillance, like humans like to do bad stuff to each other. And we're going to use these new tools to do bad stuff. Now, is it an existential risk? I don't know. Could we build like the super virus that kills us all?

with the help of AI, maybe I don't actually know, how easy it is to control the inputs to that. I think that worries me the most is bio-weapons. Whether it's created by a human or an AI agent, I actually think that human is more likely because we know humans are terrorists. We don't yet know how to make an AI be a terrorist, but we know how to make a human a terrorist, just let them go, like humans do it already. Right. Well, if a human could become a terrorist that could kill people on an existential scale with the help

of their robot friend. Correct. I am worried about that. Yeah. This is a smaller worry, but, you know, it sticks with me, especially because of what we've just been through with social media, which is I worry about all the ways that AI and specifically AI chatbots could replicate and amplify some of the problems that social media has fueled,

like mental health challenges, especially among kids and teens, this trend towards loneliness and isolation. And so sometimes it worries me that like our sycophantic robot friends end up replacing actual human relationships. It's possible. You know, humans are inconstant. They're, you know, like, moody. They are doing their own thing.

They're selfish. They change over time. They, you know, are sometimes assholes. Like, you could imagine building perfect companions of robot friends. And if you anthropomorphize your robot friends enough. I mean, like, we may even anthropomorphize each other. Like, sometimes I think that humans, you know, our own relationships are us anthropomorphize. each other, just sort of making up stories about what each other are like so that we can get

along. We interact with these constructs of other people that make in our own minds instead of interacting with a person as they really exist. And we could do that to robots. There's no reason we can't do that to like a chatbot or something. Yeah. We are machines, our socialization is built around modeling other people. We build models in our heads of what other people are like. We don't really know, you know, you don't know what it's like to be me. I don't know what it's like to be you. I'll never know. We'll never know. Instead, we make models of each other. I think, okay, if I were a guy who had this job with that kind of shirt, that haircut, and maybe I know a little more about you, where you come from, the more I know, maybe the better

model I can build, but it's still just going to be a model, right? There's going to be tons of stuff that leaves out. It's a simplification. Yeah, but the more you interact with someone, the more you get to know, there's plenty of examples of married couples have been together for decades and decades that, like, didn't know some huge thing about the other. That's true. I guess I worry about, I mean, look, there's mental health challenges here, but, and you've written about this a lot. Like when I think about social media and some of the problems of social media have fueled, if not created, it's made it more difficult, I think, to practice democratic governance, to like live in a democracy with other people. And it has given us the sort of illusion

of connection with one another. And in reality, I feel like it has helped make us even more divided from each other and angrier at each other less trustworthy of each other. So I, you know, I do worry about AI sort of supercharging that. Yeah, I worry about it too. But, but I also think we always fight the last war. Like, we weren't worried about the internet doing this, right? No. Why not? Because ultimately, TV and radio didn't do that. Media we'd had before, or communication devices, the telephone. None of these communication and information devices, that we'd made before did that you know sure radio was used by like fascist to organize some of their fascist stuff but like ultimately radio is pretty benign yeah like ultimately tv is pretty benign

we had all these worries about it it's pretty benign so we're like okay the internet's gonna be great gonna hook us up all together our world would be better you know linked up more because television radio print and all these other new media technologies that always seemed to have a positive effect and so we weren't ready for what was coming because we fought the last war but this time I think we're going to fight the last war, too. And almost all of the discourse around AI is actually people realizing the problems with the Internet and taking this out on AI, just like all the sort of positivity about the Internet, a lot of the positivity of the Internet was people feeling good about how TV and radio and whatever

turned out. So you think there's a good chance that AI could help solve some of the problems that social media has fueled? Yes, I think there is. I don't, I'm not telling you this will happen. Right. I'm not saying this is what's going to happen. I'm not telling you that I think it will happen. It might be useful to think about some scenarios of how AI could not be like the internet. And I think the easiest way to do that is to think of scenarios where the AI could actually correct some of the problems we had from the internet. So for example, everybody yelling, yelling, yelling, and social media, we have very good evidence by now that social media elevates and promotes the shoutiest, most divisive people. What if AI can shout down those people? And, you know, What if AI algorithms can gatekeep those people out of the discourse? Just like humans used to do like at CBS, right? I'm sure the CBS guys weren't angels.

You know, I'm sure that they were like, I don't know, like secretly fascist or communist or like having affairs doing drugs. I don't even know what they were doing. But the point is that like they were pretty efficient gatekeepers in that they kept a lot of these crazy shouters out who are now just dominate Twitter and podcasts. They kept them mostly out, right? We didn't have Nick Fuentes and Hassan Piker and all these assholes running around. saying like, you know, America's evil needs to be brought down, kill all the Jews.

You didn't have that shit. How would AI potentially do that? Like, is that just like downrank? I mean, Elon Musk could do this right now on Twitter with some algorithmic tweaks, right? You get 100%. It's not so simple as a push button thing because like people will see it's getting done to them. You have to make it subtle. So you think AI could make it more subtle?

AI could make it more subtle. I think eventually AI will figure out how to gatekeep. And whether we gatekeep by promoting like Nazis or whether we gatekeep by promoting like Nazis or whether we gatekeep by promoting reasonable middle of the road people depends a lot on who owns the I companies, who owns the algorithms, how does the government get involved, blah, blah, blah, you know, for the last stuff, remember for newspapers, for print, the people who initially owned all the newspapers were some of the worst people. And if you look at like William Randolph Hearst, right, Citizen Kane himself, like that guy, he like promoted the most racist, jingoistic militarist bullshit that said we've got to kill all the Asians, kill all the, you know, whatever, whoever. War, race war, the white race must, you know, kill all the Asians. What a fucking asshole, man. I didn't like that guy.

You should stay playing with the sled. And like, I mean, they made a whole movie about how that guy was an asshole, but that was bad. Eventually, we got these, like, very, you know, we developed these norms institutions by these monopolies that, like, the New York Times was kind of a monopoly company or the Washington Post or like, all these things were local monopolies because they had like efficient distribution networks for physical papers. and efficient networks for getting local classified ads. And that was their monopoly. Local monopolies, not national. And so they developed these norms around good journalistic practice, you know, what it meant to have a newsroom, what it meant to cite a source and all these things,

that William Randolph-Hurst couldn't care less about. Right. That was developed over time, and we learned to trust those things, and it took us hundreds of years. I mean, if you look at their first things that were printed, it was basically like all just like genocide and murder advocacy like the French Revolution was was totally motivated by like all these pamphlets and broadsheets and stuff like that

that were like you know the rich people have taken everything from you you must kill them in their sleep you know it's like that's what they were doing it was very Twitteresque yeah right it was very much like Tucker Carlson's podcast but worse and so like that's what happened it took us a long time to sort of evolve defense mechanisms against that in society so that we had reasonable people in charge eventually we sort of did

not everywhere did some places just had government propaganda but you know some places the newspapers are relatively reasonable they sometimes told lies they sometimes believe fake news but then at least it was more benign than probably what we've got now which was ruled by the fakesst and the evilest and so like AI algorithms eventually over time they could become the replacement for like the guys at CBS back in the day Walter Cronkite digital Cronkite I think the people who run the companies are receptive to this and they don't agree on what that should look like, but I think they understand now that they are going to become in the position of William Randolph Hearst. Where I see this already playing out is like, you know, GROC is quite different as an LLM than your chat GPs and your clods and the rest of them,

just in the way that it sort of mimics some of Elon's politics. And so I do think that like so much of the hope of, AI is dependent upon the people who run these companies, and I'm not sure that I have a ton of faith in those people right now. I don't either, but I didn't have faith in William Randolph-Hurst. So for a little while, we got used to thinking of the press as like this participatory anarchy. That was more like in the days of the French Revolution or the 1700s. Like the press was a participatory anarchy. You could just have your own little printing press in your house out in the country and print like broadsheets and pamphs.

Let's distribute them. That was like the analogous to the social media discussion age, but it didn't last. Now we're going to have to start thinking of the press as the people who control the algorithms, the AIs that create and curate the algorithms that give us your algorithmic feed. That's the new press. Elon Musk is William Randolph Hearst now and Dario and whoever's running Gemini, I guess. Like those people are the William Randolph Hearsts of our time. I mean, there was never just one William Randall first.

There was a bunch of them. But, like, that's how we have to think about the press because that's who's going to control what you see. There was someone who controlled what you saw in 1965. There were some guys in a back room who decided which facts of the world, which information you would see and which you would not see. In every country, in every system, someone somewhere was controlling what you saw because TV had economies of scale.

And that's why they did it. And radio to some extent, but less. and print to some extent, but less. All those things had economies of scale. Economies of scale led to industrial concentration. Concentration created space for someone to control what you saw. And that stabilized society because industrial concentration was what enabled gatekeeping. And gatekeeping was what kept out the crazies with the broadsheets saying kill them all, kill them all.

And we're going to get that again. We're going to get the gatekeepers back. We're going to get the industrial concentration back. I think because I think feeds, algorithmic feeds have still have some sort of, you know, network effect of some sort, like the more people watch TikTok, the more people make TikTok if humans are the ones doing the creation. I guess we could have AI slop, do it, and then we have to ask what kind of economies of scale AI has. Anyway, there's multiple blog posts in here to be written in the future. But like, I guess the point is that we're probably going to get gatekeepers back

in some way. We're not going to have the anarchy that we had in the 2010s, where it was just like every shoddy schmuck seizes the bullhorn and goes viral. Yeah, the gatekeepers go away. Everyone has a bullhorn and there were obviously problems with the gatekeepers. It does feel like, you know, the pendulum could swing back a little bit towards gatekeeping. But if it swung back towards gatekeeping where the gatekeepers are, just imagine, you know, all Elon Musk's with his politics or like all William Randolph Hearst. Right. Then you get Pravda. I want Pravda. Is there, and there is their room for other voices, for other perspectives and viewpoints with if the, if the, if the, the, if the, the technology. Is this powerful?

I don't know. I think, unfortunately, the shape of technology determines the balance between independent and gatecapped. And that ultimately, that balance has to be a balance. There's some sort of internal solution, that some sort of equilibrium point that's the best, some internal optimum, right? We need to find that, but I can't find it for you.

I can't tell you where that is. Offline is brought to by MintMobil. You know, you don't have to let big wireless and your overpriced phone bill suck the joy out of the holidays this year. Because right now, all of MintMobil's unlimited plans are 50% off. You can get three, six, or 12 months of unlimited premium wireless for $15 a month. It's their best deal of the year, and it makes it real easy for you to give your expensive wireless bill the Scrooge treatment. MintMobile's best deal of the year is happening right now. Get a three, six or 12-month unlimited plan for $15 a month.

All Mint plans come with high-speed, data and unlimited talk and text on the nation's largest 5G network, you can bring your current phone and number over to Mint, no contracts, and no nonsense. Crooked Media staffer Nina locked in her savings with a 12-month plan for Mint Mobile. She says the service is excellent and couldn't be happier about her decision to switch to Mint Mobile. Turn your expensive wireless present into a huge wireless savings future by switching to Mint. Shop Mint Unlimited Plans at MintMobile.com offline. That's MintMobile.com slash offline. Limited time offer, upfront payment of $45 for three-month, $90 for six-month, or $180 for 12-month plan required, which is $15 a month

equivalent. Taxes and fees extra, initial plan term only. Over 35 gigabytes may slow when network is busy. Capable device required, availability speed, and coverage varies. C-Mintmobile.com. I want to talk about social media since you've written about that a bit, and obviously social media made us trust each other less angrier. Your case is to why. that happened is one that I hadn't heard before, which is that social media took us out of our ideological bubbles and forcibly exposed us to opposing views, basically throwing us all in a room together. Can you explain that a little bit? I thought that was interesting. Yeah. So like in

2005, suppose you wanted to live in a conservative place where Christianity was the norm and, you know, traditional, everyone respected traditional marriage and all these things. You could just move to, I don't know, Oklahoma. And you could live that life. You could have that life. And if suppose you wanted to live a progressive existence where people like did land acknowledgments and, you know, everybody composted, you could move to San Francisco. You could do that. And the people you'd be surrounded with would be people who were generally like you. We could sort, right?

Then one day there was Twitter and the old Facebook news feed. And then that day, all those people who had spent decades and decades sorting themselves into like-minded communities were all thrown into a single room with each other. instead of the internet being an escape from the real world, the real world became an escape from the internet. And as Jean-Paul Sartre once famously said, hell is other people. We were unable to sort ourselves into like-minded communities. So imagine that Steve Miller got his way and we deported all the immigrants back to like Somalia or whoever, right?

Yeah. We poured all immigrants. Stephen Miller gets his wish. They'll still all be there. They'll just be online. And because online is now where you live, you won't have gotten rid of them. the only country that has actually gotten rid of the foreigners is China because they built the great firewall and so you can't talk to the foreigners.

China is the only country that's preserved localism because they used press control to force it. Right. Media control. And so, but the fact that we all live online now and that online we're exposed to foreigners constantly means that you can't get rid of the foreigners by just physically kicking out the physical foreigners because you weren't talking to those people anyway. If you lived next to like some Somali family, maybe you'd talk to them, maybe you'd like have them over for dinner or whatever, but you don't have to. And most people don't. Like,

people in America don't know their neighbors. They haven't since we got cars because we're so mobile, right? We don't know our neighbors anymore. It's not because of immigration. People don't know their neighbors in like any, in any country where like you have cars and mobility. People don't know their neighbors. And then so we're all just on the internet all the time. So like kicking the Somalis out of your neighborhood or whatever or the Afghans, the Haitians, whatever, whoever you're going to kick out, it might have some effect, but it's not going to affect your soul. social life. It's not, you know, your social life is online now. You're sorting into like-minded communities online. And whoever's there, whatever people are like, you know, you're interacting

with there, though who gets physically removed from your neighborhood is not going to have a big effect on your social circle anymore. Because it's online. And so I think that's what happened. Our country stopped being able to spread out and sort geographically because geography suddenly stopped mattering. That's a oversimplified story. But that's, That's basically what happened, I think. And then it just made us angry because we were hearing opposing political viewpoints that we necessarily wouldn't have heard otherwise because it's not like our neighbor that we didn't know that well, would just like walk out of their house and start like

screaming about, you know, how the, whatever you hear online. Exactly. So like everybody lived in the town square all the time. What kind of society is that where people all live in the town square all the time? Communal bullshit. Like, you know, imagine in 2005, if you lived in a red state, how often would you have people say that, like, all white people are racist? You'd never hear that. No one would say that to you.

But in 2014, you'd see it every time you opened up Twitter. You'd see someone saying, and they didn't live next to you. They lived in California. They lived in New York. But they'd say it and you'd see it. And then you'd think, wow, my society is so intolerant. Brooklyn was always like that, but you didn't hear it because there was no way to. for you to hear it. People geographically sorted. Then one day, you heard it all. Or like, suppose

you're living in San Francisco in 2005, and San Francisco's created this gay-friendly community, you know, maybe in the 70s you'd hear gay people suck. You know, that's unnatural to her. You know, maybe you'd hear that like in the 70s, but then San Francisco became this very gay-friendly place where you could just go, you could be gay, no one would bother you. It was just normal, right? Then one day, in like in the 2010s, when you get Twitter and stuff like that, suddenly you're hearing a million homophobic people yelling at you. Like, I can count the number of times I experienced anti-Semitism before social media, mass social media.

Like, I can count it on my nose because once I met one guy who was anti-Semitic one time. I was very surprised. Growing up, this Christian conservative place in Texas never once. Never once. I never heard anybody say anything bad about Jews. Like, I knew that, you know, I could read about the Holocaust like a book or the Spanish Inquisition or some shit. or like how like Jews were like discriminated against in the 50s and like Princeton but like it wasn't

real it wasn't in my society I never encountered it and then then one day on Twitter there were like hundreds of people saying gas the kikes race war now every day I'd open it up and I'd see those guys that was jarring that was new and it was because of social media everybody went through that everybody had someone out there in America who hated the shit out of them and everybody woke up one day and met that person. I'm surprised our society held up as well as it did. Yeah. Wow. So do you think we're moving away from that now? Yes. Yes. We're learning our lesson. We're separating into like discords and group chats and offline. Mass social media is dying. People don't go on Twitter that much anymore. People stare at algorithmic feeds. They look at TikTok.

They look at scroll videos. We're still in touch with each other, but add or remove. We're making TV for each other. It's TV. There's this impersonality of TikTok where like, yeah, you can make a TikTok response to what someone else says, but you're not talking directly to them. It's not like back and forth in real time. And it's mediated by an algorithm too. So if people don't want to see that politics shit, like if I, TikTok is going to pretty quickly realize that I don't want to see anybody making a video that says gas the kikes, race war now. You know, like they're going to realize, oh, you don't actually want that. So I'm not going to show it to you. And then algorithms put us into our bubbles. And that's good. People are oh no you're going to be in a filter bubble you're never going to see anyone who disagrees

with you and then you see the actual person who disagrees with you and it's some dipshit saying gas the kikes race war now that's who just that's the opposing view do i want to see the opposing fucking view all day no thank you i would like to talk to people who are somewhat on the same page as me and can maybe debate stuff with reasonable modest disagreements you know who aren't just like screaming for me to die yeah point taken on that for sure it does feel like living in in a big multiracial, multi-ethnic democracy of over 300 million people, we do need to figure out how to interact with and debate and mediate conflicts with people who are very different than us and have opposing views. We just still need that muscle, right? Yeah, we do. But you know what?

Do you think we got that muscle in like 2005 or we just, we were in these homogeneous filter bubbles? Well, yeah, I was going to say in 2005, it was better than the heyday of Twitter. When was the best? When was the best for debate? The best period of our history? I don't know. Like early 2000, I don't know. Well, I would compare, I feel like from 2015 to now has been certainly the worst of my lifetime. It's shit. You wrote a few pieces about America's identity crisis and the opening for a patriotic defensive, traditional liberal American values. And, you know, it's an opening that I think in some ways was created by the, you know, the MAGA post-liberal right, also the far left, and one that the Democratic Party, I think, has at least to be as generous as possible, not taken advantage of just yet.

What would that kind of liberal nationalism look and sound like coming from a Democratic candidate, say, for president in 2008? I would like to point out that Kamala Harris improved mightily in 2024. They had a bunch of like pro-America stuff at the convention. It didn't make it through because this stuff has to be repeated and repeated and repeated over like a decade, I think. Yeah. But it was a damn good start.

And we haven't seen this stuff from Mamdani because Mamdani is allied with like the left, which is like hypercritical of America. And they're still running off the like gas fumes of like Soviet propaganda from like a million years ago. So like the Zoran, you know, Bernie left, like they're not going to do it. They're like structurally anti-American. But then regular progressives can do it all the time. Like people who are not allied to like that small faction within the Democratic Party can just do this very freely. And so, you know,

we saw it with Obama. Obama did this incredibly articulated like the goodness of America in a way that liberals responded to and that normal people could believe. He did it really well. Just go back to that. Like just look up old Obama speeches and do what Obama did. There's nothing like the world hasn't like moved on in a way that makes it impossible to do that. Just go back

to that. Yeah. Just like look up fucking Barack Obama. Like he's great, you know? I like to think so. I've been waiting for this for some Because I think that sometimes Democrats, moderate Democrats, talk about my problem with them, which is that they have the policies that align with the broader electorate and they know the politics, but sometimes they come off quite boring and quite, and I feel like you do need sort of that, there's something more to politics and to being a leader and to being an elected

politician than just figuring out the right economic policies. But it is speaking to something bigger about the country's identity. You've been talking about America's identity crisis, and I do wonder if in the next couple years, as we head towards 2028 and beyond, if a Democrat's going to be successful or the Democratic Party is going to be successful, they're going to need to have that sort of bigger story about America, defensive American values that, you know, I think that the left or the broader coalition of the Democratic Party can see itself in if they want to be successful. Yeah. A lot of progressives got really drunk on the idea that, like, there was. this invincible like demographic drift toward make like you know as our country becomes more Latino we would become more democratic and Trump has disabused them of that I think they understand now that Hispanics can be very conservative and vote Republican and it will be willing to vote Republican and I grew up in Texas so I knew that already but they did not because they grew up

in New York and California where you know like Hispanics are just as progressive as everybody else. So they thought that was a structural thing. They looked at the national statistics. Like, oh, Hispanics vote 60, 40 Democrats. So just get more Hispanics and we'll win. Look what happened to California. But it doesn't work like that. Yeah, no. You need to go out and actually persuade people. Can't take anyone for granted. Um, Noah, thanks for joining offline. Awesome. That's been great. Did I manage to convince you of anything? I feel a little bit better about AI. I do. I want to join because I'm like, you know, I was getting towards AI Dumerism. And I've had had a lot of people on who were quite negative on AI.

And so I thought we needed a different perspective and we got it. And I feel like you made some good points on that. Cool. All right. Well, thank you very much. Appreciate it. Absolutely. Before we go today, some quick housekeeping.

If you're not sure where to get your family and friends this holiday season, think about giving them the gift of content. It's the gift that keeps on giving. A friend of the pod subscription will keep them informed, entertain, and mostly sane. They'll get access to all the good stuff, exclusive content like Polarcoaster with Dan fight for ad-free episodes of all their favorite Crooked shows and endless bragging rights that they support independent progressive media. Now that's a present. Gift a friend of the pod subscription, grab one for yourself, or learn more at crooked.com slash friends. As always, if you have

comments, questions, or guest ideas, email us at offline atcrucid.com, and if you're as opinionated as we are, please rate and review the show on your favorite podcast platform. For ad-free episodes of Offline and Pod Save America, exclusive content and more, go to cricket.com slash friends to subscribe on Supercast, Substack, YouTube, or Apple Podcasts. If you like watching your podcast, subscribe to the Offline with John Favreau YouTube channel. Don't forget to follow Cricket Media on Instagram, TikTok, and the other ones for original content, community events, and more. Offline is a Crooked Media production.

It's written and hosted by me, John Favreau. It's produced by Emma Ilich-Frank. Austin Fisher is our senior producer. Adrian Hill is our head of news and politics. Jerich Centeno is our sound editor and engineer. Audio support from Kyle Seguyen. Jordan Katz and Kenny Siegel take care of our music. Thanks to Delon Villanueva and our digital team

who film and share our episodes as videos every week. Our production staff is proudly unionized with the Writers Guild of America East. I'm going to be. Thank you. Thank you.