Offline with Jon Favreau - What Is A Human For?

Episode Date: June 27, 2026

AI is creating a crisis of agency where people are becoming paranoid that they’re being manipulated, suckered, and ultimately replaced. Charlie Warzel joins Offline to elaborate on his Atlantic essay, “The Feeling of Control Slipping Away,” which illustrates the myriad ways AI is driving people insane. He and Jon talk about whether human creativity is endangered, if AI is anything more than a corporate black-box, and what it means to be human when a robot can do everything better than you. They also discuss Elon Musk's SpaceX IPO, what its valuation means and Musk's efforts to rewrite DOGE's devastating legacy, as well as his own. For a transcript of an episode of Offline, please email [e-mail].

Offline is brought to you by Helen Keller International. It's easy to feel like the world's problems are too big for any one person to solve. But Helen Keller believed that progress happens when people come together. She famously said, the welfare of each is bound up in the welfare of all. That's the idea behind Helen Keller International, a nonprofit celebrating their namesake's 146th birthday this week. In 20 countries around the world, they work with local partners and governments to improve nutrition,

protect sight, and fight disease using cost-effective evidence-based solutions. To celebrate Helen's birthday, every gift from now through June 30th will be matched, doubling the impact supporters can have. A gift of $146 when matched can provide medication to prevent and treat neglected tropical diseases for more than 200 people, or can provide vitamin A, a nutrient essential for protecting health in sight for 146 children. If you're looking for a meaningful way to honor Helen's legacy, visit Helen Keller, I-T-L-org slash offline, before June 30th to make a matched gift. That's Helen Keller, I-N-T-L-org slash offline. When it's time to scale your business, it's time for Shopify. Get everything you need to grow the way you want, like all the way.

Stack more sales with the best converting checkout on the planet. Track your chichings from every channel, right in one spot, and turn real-time reporting into big-time opportunities. Take your business to a whole new level. Switch to shop. Spotify. Start your free trial today. Aiden Walker wrote this, or actually it was on a TikTok, but has this theory. It's called the cuck theory of the internet, right?

So funny. Which is just that like when you're watching TikTok or an Instagram reel of like, you know, an AI cute animal or whatever, or just something that's like generated in that way, like what you are actually doing is you are sitting in the in the cuck chair and on the bed is an algorithm that is trained off of all of this artificial intelligence data at the same time and your data and everything. And this thing also knows what I like because it's fed off of everything that I've ever done. And I'm just going to sit here and let a little drool come out of the corner of my mouth as I enjoy whatever this is, is a fleeting pleasure.

I'm John Favro, and you just heard from today's guest, Charlie Worsell, host of the Atlantic's Galaxy Brain podcast. Charlie's been a frequent guest on this show. I think last time we talked about Maltbook, the AI-only social network. We've also had quite a few conversations about Elon Musk. I invited Charlie back this week, partly because there's more Elon news to discuss. His IPO of SpaceX just briefly made him the first trillionaire of the world has ever seen, but also because Charlie just wrote an incredible piece about what AI is doing to us, which is slowly but surely robbing us of agency.

Charlie and I talked about how AI is reshaping what it means to be human already, how a recent writing prize went to a piece that seems to have been written by AI, how AI strips our agency from us, and the growing movement to decouple technology from algorithms. Charlie also watched me through how Elon's SpaceX IPO has been mostly built upon the value of Elon's ability to capture attention and tell a story, how financial markets are detaching themselves from reality. We'll get to that conversation in a moment, but before we do, if you're a friend of the pod, thank you.

If you're not, please consider this your midterm reminder. here's what you get when you subscribe. You get ad-free episodes of all your favorite pods. You get access to our excellent substack newsletters, including open tabs. You get Dan Pfeiffer's Polar Coaster podcast, our extra episode of POD Save America called Pod Save America Only Friends. And you get to support proudly pro-democracy independent media, which is always great. So head over to crooked.com slash friends to learn more and subscribe today. All right.

Here's Charlie Worsell. Charlie, welcome back to Offline. Thank you for having me. You wrote an Atlantic piece last month. I've been thinking about the feeling of control slipping away where you argue that the flood of AI is creating what you call a crisis of agency, that whether or not we personally use these tools, all of us are sliding into a more passive, reactive role.

We're sort of sitting back while the machines do the writing, the searching, the deciding. I want to talk about all that, but first, like, how have you noticed this in your own life? Like, where has it felt like control is slipping for you? Well, I see a lot of people, I mean, I'm a writer by trade, and I see a lot of people talking about the fact that, like, there's a big conversation on Twitter right now that is, like, roiling about, like, yeah, there's not going to be any writers going forward. Like, you know, everyone is going to eventually just give up their writing about. to these LLMs or like writing is going to be like vinals, right? It's going to be this outmoded

thing that like a couple of hobbyists have or like an elite mode of communication that just, you know, these princely people in mahogany rooms with leatherbound books, you know, talk about writing. So there's a little of that where people are just like, like give up on your stated profession, the only marketable skill that I actually have. So there's a little bit of that happening. But I think broadly speaking, this idea of agency and the kernel of it where it came from was actually this scandal, or not scandal, but this viral marketing fiasco where this company was accused of planting all these audio clips on TikTok for the band Geese, this sort of polarizing indie rock band that has gotten very famous over the last couple years. And the idea behind all of this

was the geese was a sci-op, right? Nobody actually likes this band and they're hugely popular because of the fact that these marketers have, you know, schemed and, you know, basically astro-turfed social media with all these like fake accounts or seating it with all this audio to make, you know, their music seem like it was more popular to the algorithm. so that the algorithm would promote more stuff from the band. And it reminded me a lot of the Cambridge Analytica stuff. You know, how everyone, when that happened, everyone, self-included a little bit, was like, oh, is this mind control, right?

Like, are people using these platforms in such a way to, you know, hit us at the exact right moment and make us think things that we wouldn't have normally thought? But that geese sci-op story to me was just a bigger, story about the the worries, how we're all just like shadowboxing this idea that the fixes in on everything. We're being manipulated in every way. Nothing on the internet is real. Anything that you like is, you know, an industry plant or some kind of bit of propaganda. Everything you don't like is out there in the world just because of the fact that other

people are trying to make it so. And I felt like that was the kernel of this, this feeling when so many people feel that everything they see is fake or suspect, that we're losing, not just our ability to distinguish reality, we're losing this ability of control. We don't know what is popular or what is not popular. And I think when you pair that with the idea that like generative AI is essentially a scale machine, right? It is an ability to generate as much of whatever thing you want. And now they have these agents that are starting to work on our behalf.

and do things, it was like, oh, we actually are in this, in this crisis of agency. Whether it's real or not, we are all feeling this, that this sensation of, you know, we don't have control over anything. I know what you're saying. And even you telling that story about geese, I'm thinking, I was like, so is the accusation that they did it? Is it an influencer firm that did it? Is it?

and what is the how do you even go about doing that and isn't it possible to just like music or not like I know what you're saying like I have also on Spotify for they had some AI generated songs that like made it

into you know pop songs of the week or rising whatever and I was like listening to it because it just like came on and I'm like oh this is an interesting song and then there was something about it that made me think I was like is this real and that it and it took me like

10 seconds of searching to realize that it was, in fact, an AI-generated song that had just become such a hit and went viral that it, like, made it onto Spotify. And I was like, this is not good at all. Yeah. Well, you know, the reason why the geese thing, you know, hit me in a certain place is that, like, I really enjoy the band. And I understand, though, exactly why people wouldn't enjoy the band, right? Like, the singer has this really weird kind of style. It can be a little bit atonal to certain people. It's music that the first time I listened to it, I was kind of like, I don't know if I like this. And it deepened over the course of my listening. But there's this feeling that I think the more fascinating part of this to me is this notion that

people are being confronted with something they don't understand and the broader architecture of the internet right now and this ability to generate synthetic stuff or this ability to, you know, plaster social media algorithms with stuff at this inhuman scale to astroturf them, it's allowed people, it's kind of like the idea of the liar's dividend with misinformation, right? If there's so much bullshit out there, you can just, you know, claim that anything isn't real by citing that, even if it is real. That's exactly what's happening now, but with culture, not just with politics and things like that, just with anything, with art, with media. Is this written by an AI bot, right? And then you have all of this real stuff,

happening, like the, you know, this possibly AI-generated short story stuff winning prestigious prizes where it's giving credence to that idea. And it's really brain scrambling, I think. Like, your experience with the stuff on Spotify, everyone is having that experience, right? And I think, you know, you see it especially with, like, jazz, right? Like most of the, if you're listening to ambient music in a coffee shop, there's like a really good chance you're listening to something that some guy in like a click farm is just like hitting a button and like popping out so that it gets put on these playlists because it's fine and generates a whole bunch of money for them. Meanwhile you have the people who are like actually playing in these coffee shops scraping together a living and we're all just like what are we doing here? Well and that's the problem is of course that most entertainment and culture is just fine including very popular entertainment and culture.

So it becomes even harder to detect. I do want to talk about the literary scandal you referenced, and you referenced in the piece as well. So this year's Commonwealth Short Story Prize went to a story called The Serpent in the Grove by an unknown writer named Jemir Nazir, published by a well-regarded British literary magazine. And almost immediately readers started flagging it as AI written because of lines like, she had the kind of walking that made benches become men.

and the girl smiled like sunrise over a sink. Then two more of the prize winners got swept up in the same suspicion, and now the AI boosters celebrated this as a triumph because it's proof that the models are that good, and the skeptics called it a slop tipping point. Which do you think it is? I mean, I think it's essentially, I think it's both, right? When I was looking into all this,

the the when I am thinking about the the literary scandal that that sort of put into focus this major question I was like what is all this agita about right and I think it's around it centers around this this one question in the AI age which is kind of wild that we're actually sitting here asking ourselves this and it's what is a human for yeah if if people who are supposed to be very literary in their own rights such that they can be the judges of prestigious literary competitions and they are falling for this stuff or they just like this AI written prose right like i i think the benches become men is a really shit line i don't think it's good at all uh as a writer myself i don't understand i i've tried to understand it i don't understand it right but if they like that

it's not just a question of oh you were duped or you weren't duped or whatever it's actually like if that writing is appealing to you on a human level and you can't see the difference between that because because you're you're blighted by that then it brings into this question what what is the human for there that like if we can't distinguish between these things if we like these things if it's good enough that it is you know that it is outpacing some of the stuff whether it's on the the literary prize level or the, you know, the jazz music on Spotify thing, then it calls into question, like, what is, what is, what are we doing with art? What are we doing with judging the merits of this?

These are like these massive existential questions that all of us, just scrolling social media or trying to find a song to play, like are now being confronted with in these really minute ways all the time, right? Well, what does it mean if I like this? And it's not real. Should I care? I don't know. And I think that is, that's really brain scrambling.

That's like, that's like, you know, the stare into the middle distance for a long time stuff. What's your answer to the question? What is a human for? Yeah. Well, I mean, I, this is. You must have turned it around in your head after writing it. Well, I think, you know, in some way, I'm just like, I'm really resentful of having to, like, think about that question on any level. Like, I think life is actually pretty difficult for most people on a day-to-day basis without having to sort of consider the usefulness of the species, right?

It's a little bothersome, right? There's so many problems we have to figure out on the human level, on the, like, how do we exist? How do we, how do we, you know, lift up the people that need help? How do we, you know, how do we figure out our government? How do we reopen the straight of Hormuz? Whatever the hell it is, right? It's like you have all of that and then you have this, this bigger existential thing. So what is a human for?

You know, I think that when it comes to the art with the writing, et cetera, I think it's, I'm compelled by the notion that, you know, when you read something that's written by another human being, it's just extremely powerful technology. It is my words, your words, whoever's words, rattling around in your brain. It's like this very intimate, unbelievable way to, like, fuse two people together or two people's ideas. I think when you think about art in other ways, like AI art or something like that, real art, photography, whatever it is, it is a way of someone taking something that has either moved them or that they think is moving in some way, elicit some kind of response. And again, going into somebody else's mind and basically like interjecting that thought into there. So it is a connection between two groups of people.

And I think that's what all of this is. I think this is why we loved the internet in the first place is like there's this crackling static of like, oh my gosh, I put I, I type something in and a whole bunch of people had a reaction to it, right? Whether it's good or it's bad, it's authentic. It's real. It's communication. Like we're put on this earth. to interact with other humans in other ways, whether it's good or bad or ugly. And now we have this third thing, this extra dimension to the whole conversation that scrambles

that. And I think it undercuts the premise of not just the internet, but also art and, and, you know, like fundamentally what we're supposed to be doing up with this one precious life we have. Yeah, it made me think because I think, I think, humanity is one of the things that makes it special is our individuality. And I, you know, as someone on the on the left side of the spectrum, I think a lot about, you know, collective action and why we need each other and why sort of social interaction is sort of the core of the human experience. And I very much believe all that. But when you have now AI, which is generating its intelligence, from the sum of all human intelligence together.

And it's like the ultimate hive mind. I mean, this is pluribus. I don't know if anyone here has seen pluribus, but it is like the ultimate, it is a show about AI, I think, even though that has not been explicit in the show yet, where it is just this, okay, we're going to take all the different knowledge in the world

and we're going to try to figure out sort of the best summation to spit out back to the person. But you're never going to get individuality from an AI, and you are going to get that from someone who is a writer, an artist, and entertainer. Ideally, if it's good, it's going to be something that only can come from that person and can never come from anyone else. So I do wonder if that is, that's the thing that's going to save us, but I don't know. The New Yorker writer, Kyle Chaka, had a column recently about how Claude design the program that, you know, a lot, like the Cloud Code program that allows,

you to design websites and things like that with the kind of like the click of a button and a prompt. He had this piece about how the design is very similar, right? There's just like across every cloud code design website, you have all these tics, visual ticks, right? I think you also see a lot of this with AI art, that there's a lot of similar ticks. The writer Max Reed wrote about like bootlegged, like NBA, finals like Knicks gear and how like it used to be this like kind of really janky you know photoshopped strangeness right that was unpredictable in a sense even if it was if it was bad it was

still kind of good because it was like one of one and like some sicko graphic design is my passion person you know is like is giving you this thing and that AI art has kind of homogenized that right because it has raised the floor but lowered the ceiling such that you don't get, you get this like very, you know, mid, not even in the sense of whether it's good or bad. It's just like it's not going to be offensive. It's also not going to be super weird. It's just going to kind of be. And I think that that in a sense is a really interesting, I just keep citing different people here,

but there's this conversation on Twitter the other day about about. that Claude thing and it said the person basically said this is my argument for why there isn't artificial general intelligence in these models right now

because a generally intelligent you know like a human type thing wouldn't have those type of ticks overall right it wouldn't just default to this sort of safe mean in the middle

it would be completely unique in these really surprising ways like it would design a website that you've just never seen before that's just whoa you know and I think that it speaks to the fact that it's trained these models are trained of the entire corpus of whatever and there's a lot like as you said about art uh earlier there's a lot of just like okay very fine very like middle of the road safe stuff and when you look at websites

everyone uses Squarespace or wicks or whatever it is right to design these things in these templates so then the templates become over indexed in the training data so then it just feeds you this stuff that is like exactly what you're supposed to see that's safe, et cetera. And so I do think that is a safe ingrace, at least right now with these models, is it, they, they do create this really mid safe thing because they're not humans, because there's no, you know, the model isn't like, like, watch this. I'm going to, I'm going to blow their goddamn minds, right? Instead, it's like, ha, like, I got to, you know, I got to get this done on deadline. I had this experience where I started out using chat chie pt and then found it so sycophantic i was like i can't and i thought it was

the i thought it was chat chept being sycophantic that really got me more than the ticks um because then i used clod and i was like oh cloud is not as sycophantic i like clod is much better i found this is the l lm i want to use but then the more i use clod it's like it's not a sycophantic but it has like chat gt has a different set of ticks and now and i use it for research But now when I ask it questions and it comes back and it's the same, it's just like, it drives me, it's starting to drive me insane. Like, Claude is starting to drive me insane just because it's saying the same fucking thing over and over every time I ask it to find something. And it's like, this is the load bearing, you know, sentence for your, that you want here and the fact that you were looking for. And I'm like, how many times you're going to say load bearing here?

It was just, it's a weird thing that it's like, it becomes, I wonder. how you improve upon that. The answer is, oh, well, they're getting faster and better and smarter all the time. But at some point, you can't get smarter than all of the human intelligence that currently exists. Right? Like, it can't go beyond that. So it's always going to hit this limit, I feel it. I don't know.

Yeah, I mean, it's really, this is what actually interests me about artificial intelligence is the way that these things iterate the way that they get better, not in just the, we're throwing a bunch of more compute at this and there you go. But the ideas of what can what can make these models create emergent properties, right? Like I was listening to the psychologist Alison Gopnik who writes and talks and thinks a lot about and studies like intelligence and especially like childlike intelligence and how we learn and stuff. And there's this whole this whole possible understanding that, you know, maybe we're training the models or the way that we train the models now, the way that we design them, is actually very much along the lines of, like, how a college student

learns or something like that, right? Like, you have someone who has this basis and then you just, like, throw the books at them or you throw all these things. But, like, children learn, especially in these, you know, really early years in a completely different way, right? Like how they become their unique, they're, you know, potentially born with a unique personality. But, like, the way that they go through the world and experience the world inform so much of how you know you become who you are how you how you reason how you think what you feel and that is such a like a tactile process it's something that a these models can't go do they can't go out and have like the physical experience of being

alive but maybe there are ways you know down the road to simulate that in some way that then changes completely what a model will be able to do and think like. But right now we're in this paradigm where I think you're right. It's it is very static. It's like we can make these incremental improvements, but because we're trained off of all of this stuff and there's so much of it in that, you know, middle ground of being very safe and predictable and boring that it ends up being inhuman in how it feels. Offline is brought to you by HIMS. If you're looking for a sustainable way to manage your weight,

weight loss by HIMS offers a streamlined approach. HIMS provides access to a range of FDA-approved GLP-1 medications, including the UGOV-P-Pen and the Wigovi-Pil, the first and only GLP 1 available in a non-injectable format. Through HIMS, everything happens online. You'll connect with a licensed provider who will determine if treatment is right for you. If prescribed your medication is delivered right to your door,

no insurance necessary, and it doesn't stop there. HIMS makes hitting your goals seamless by offering access to 24-7 messaging with your care team and in-app lifestyle and nutrition support like recipes, meal plans, fitness videos, sleep content, and more. Even better, with a range of affordable GLP-1 options, HIMS makes it simple to find a weight loss approach

that fits into your world and your wallet, if eligible. You'll get a treatment plan personalized to you and unlimited dosage changes as needed. Ready to reach your goals, visit HIMS.com slash offline to get a personalized affordable plan that gets you. That's H-I-M-S-com slash offline,

Hymns.com slash offline. Based on advertised cash price for 30-day supply of medication only, membership required, fee not included, and billed separately. Weight loss by Hymns is not available in all 50 states. Wagovi is the registered trademark of Novo Nordisk AS. To get started and learn more, including important safety information, Wagovi clinical study information and restrictions, visit Hymns.com.

Offline is brought you by Sundays. If you've ever had a dog who's picky when it comes to food, you know how surprisingly stressful it can be. You'll try every brand under the sun in hopes that your dog will finally find one they like. But maybe the answer is simply that your dog knows the difference between processed brown pellets and real food. Sundays for dogs is different. They start with 80% all natural meats and then finish with superfoods like kale, ginger, and blueberries. All of the ingredients are gently air-dried instead of using high heat. Sundays doesn't look or smell

like traditional store-bought dog food. It looks like high-quality jerky the kind humans would eat. No fillers, no nutritional blends, no chemicals, just simple, complete nutrition. Founded by Dr. Tori Waxman, Sundays was created to meet her high standards as a veterinarian, and and as a dog parent. And the best part, Sundays requires no fridge, no freezer, no prep, no mess. You get the quality of a home-cooked meal

with a scoop and serve ease of kibble. It's what dog food should have been all along. I got Leo here with me today. Leo, what do you think of Sundays? And he's nodding. He's nodding. He loves Sundays for Dogs. Huge fan.

Runs right to the bowl. As soon as I pour it in, it's easy to store. He likes it. Keeps them healthy. We're big fans. Make the switch to Sundays. Go right now to Sundays for Dogs.

com slash offline 50 and get 50% off your first order or you can use code offline 50 at checkout that's 50% off your first order at Sundaysfor Dogs.com slash offline 50, Sundaysfor Dogs.com slash offline 50 or use code offline 50 at checkout. There was a New Yorker piece on the serpent and the Grove scandal where a literary critic pointed out that maybe AI isn't corrupting literary taste at all. Maybe it's exposing it. The idea being that if a machine can nail the exact notes that win a prestigious prize, the problem might not, the problem might be actually the formula and not necessarily the machine.

And so the flip side of all this is like, if AI is just handing us back our own cliches and our own house style, maybe that's an indictment of us as well. Like there was this whole conversation in The New Yorker with all these literary critics where they're like, well, you know, it does, this, this short story that won this prize does seem like the stereotype that the Western world has about developing, writers from developing countries from the global south that of course that like if you write like this then the literary critics in in new york and london might be like oh yes that's amazing and i have wondered that with some of this ai stuff is it is it sort of holding up a mirror

to some of the most cliche or stereotyped um sort of views of what constitutes really wonderful art and culture and entertainment. The thing that I like about this artificial intelligence moment, even though I resent having to think about what I'm actually useful for as a living, breathing, hunk of meat walking through the world, is forcing humans to ask a lot of questions about how we work, right? I think that that's cool. Like I really genuinely think that it is cool that there are neuroscientists and people out there who are like, you know what?

We actually, there's so much we don't know about consciousness. There's so much we don't know about intelligence. There's so much we don't know about, you know, why our brains do this thing. Like when some people were arguing very early on, these LLMs are just fancy auto-complete. And then there were other people countering with like, well, aren't humans fancy auto-complete? Right? Like, I know the next thing to say in my head to make you or the audience think that I'm smart or hopefully, right? And so that my brain is going to like launch those words down there. It's a predictive text thing based off of my training data, which is being alive and reading crap and all that stuff. And so I think like those parts are very interesting when it does like you say, hold up this mirror or make us ask these questions that aren't as existential but are a little bit more like, why do we read? like that. Why do we evaluate that? Why does that make sense to us? I think it's funny that the house style of AI text in a lot of ways that people really hate, especially across social media, is what

a bunch of us at BuzzFeed back in like the 2010s used to call broitry on LinkedIn, which is just like these like single sentence, you know, like tone poems about efficiency in the workplace or whatnot, an ROI. And that, there's so much of that out there that this has been trained on that now it's become like this house style and we hate it. And it's like, right. Well, you know, AI didn't come up with that. A bunch of gibronies did. Well, it's now like it's gone so far in the other direction that like anyone who uses an MDash, people are like, this is clearly AI wrote this. And it's like, I don't know.

A lot of people are using M dash as long before there was JETGPT. Get off my block, man. Well, it gets back to your point, too, though, about how it's also confusing and you do feel this loss of control because you're like, am I reading? Did the person who posted this write this? Did they get an LLM to write this? If I can't tell the difference, does it make a difference? Yeah. Like, it really fucks you up.

I did a podcast a couple months ago with Max Spiro, who is the co-founder of Pangem, which is the sort of preeminent AI detection software. That is by their standards and by a lot of people testing it really, really, really good. There's obviously false positives, you know, some false negatives that do come out there. But it is good at detecting AI writing. What has happened, though, is they've built these tools, including these, like, you know, Chrome browser plugins that allow you to basically, anytime you're on, like, a social media platform or, you know, someplace, it just automatically scans it and gives its, you know, its diagnosis, like, in,

your browser as you're doing it. And the thing that I was trying to pose to him and, you know, I don't know that he addressed it that well in my eyes is, aren't you creating this weird arms race, right, of people who are, you know, basically trying to, you know, throw errors into their AI stuff to make it look more human, but also this contributing to this feeling. Like, you've seen a lot of people using PANGram to detect Hunter Biden's posts on on X. I was thinking about his right when, right when I just asked that. And people just being like, well, this is, you know, 100% AI generated, like, you know, talk about inauthentic, et cetera, et cetera. And that is potentially an easy target and a believable one potentially for some people.

But there's been a number of instances of, you know, text from the Pope's. social media account and the Pope is you know very prominently against you know the inhumanity of these artificial intelligence tools and they're coming up in Pangram as 100% AI generated and now you're seeing people be like

are we are we sure that this this is right because to them that would be this big betrayal and I'm not here to speak on whether or not you know the Pope's social media team is using Claude to to write anything but it's part of this right because

it's easy if you dislike a person to say, yeah, see, like, that's fake. But if it's someone who you respect, whose, you know, whose views on all of this, on, on the humanity are really core and important to your own understanding of, you know, of, like, if they anchor you to the world in some way, and then that is a betrayal, then it just becomes, you know, what are we going to do here? But it's also, I mean, as someone who used to work in politics, there is in every profession, in industry, a style of writing, especially professional writing that's going to be public, that seems a little more bland and reminiscent of other types of writing that you've seen that's different than like your own personal writing, right?

Like if you're writing something yourself, you feel like it's going to show your character and your own ticks and whatever else. But like, you know, you're writing a statement for a politician. That could come out of the mouths of a host of politicians, any one statement. And so you'd think that maybe whoever's writing, maybe the Pope's not writing his tweets, whoever's writing his tweets is writing tweets like they would put out a press release from the Vatican, and then that's getting picked up as AI. And not to, and maybe it's not, maybe the Pope's just churning out AI written tweets.

Who knows? I don't know either. It goes back to this, like, maybe it's exposing rather than fueling this style of writing and communication that is a little too hive mind anyway. But also, you can complicate that. I think that's potentially true, right? But you can complicate this in another way, too, which is, what if there's a problem with one of, I'm not accusing this of being the case, but what if there were to be hypothetically a problem

with one of the new model updates? Because Pangram is an artificial intelligence program. itself, right? What if there's just a problem and they fix it in like three days? But in that period of three days, there's just like such a higher rate of false positives or whatnot that, you know, people are just branded a certain way, right? So I, you know, we are, in a sense, what's funny about this is, is when I was talking to Max, the founder of Pangram, he's really worried about all this AI stuff and the slop and in the inhumanity of the internet. He is also building an artificial intelligence tool himself, right?

So he is in a way also subject to all of the same issues in terms of the training data, in terms of the hallucinations, in terms of all of these concerns. And so it's a very, like if you think too hard about it for a while, you just kind of spin out and the easiest thing to do whenever reality gets blurred in such a way is to, is to retreat. It's to say, okay, you know what, taking a pause. Everything's fake. Everything's real. I don't care. I'm going on vacation. See you later. Well, just to talk about the work aspect of it, there's two phrases you pull in the piece from Silicon Valley.

One is the permanent underclass, which they admit AI might create. And the other is the only way to avoid becoming part of the permanent underclass, which is by being what they call high agency. Can you talk a little more about those two terms, the permanent underclass and high agency individuals? Part of why I feel like we're in an agency crisis is because a lot of the people who are building these tools are obsessed with the idea of agency. Being high agency is sort of the top of the food chain in terms of the new world that we live in, right? High agency people can, quote, just do things. Now, you know, that often is just a side for like, I'm good at raising money or I have a lot of money or, you know, I'm eminently employable because I started these two companies and the other ones can fail and whatever. But the idea of it is there's so much at your disposal right now. Just go make something, go do something. Like, you know, don't listen to authorities. Don't play by the rules if you can't, if you don't have to. And, you know, And they are talking about you, like you need to be especially adept at harnessing these machines and these tools. These, you know, Claude code will help you by working while you sleep be extremely high agency, which will then allow you to have more power, more stuff.

And you need to go get it now, because if you don't do it now in this little period before artificial general intelligence takes over and we have no agency, then. you're going to be part of what is called the permanent underclass. The permanent underclass is just this phrase that I think they thought was fun, that now is coming back to bite them in the ass, because there's all these people who are protesting data centers and just extremely mad and just hate artificial intelligence in every way. It polls lower than Donald Trump, and everyone in Silicon Valley is curious why,

and it's because they've been saying things. Like, you're going to be part of the permanent underclass. Like that conjures a picture in the brain of like being chained up and working in a salt mind. Like it's not a great image. You know, like Steve Jobs wasn't like if you don't buy this phone, you're going to be a surf forever. Like it was like, this phone rules.

And even if you do buy it, you still may be a surf forever. Right. Well, yeah. That was a whole other ball of wax. But it's like it's such a wild brand of marketing. But it is this idea, again. Like Silicon Valley at some point pivoted very hard. I don't actually know when it was because in the 2010s, at least the early 2010s, you know, the Web 2.0 heyday, it was very like, this is democratizing technology.

Everyone gets to say, this is going to help, you know, this is cool, this is fun, here's a bunch of gadgets, here's some apps. at some point there was this pivot I think you know like crypto was one that was very big on this where it became like FOMO became the way that you market these things get people to be very scared that they are about to miss out and I mean it's kind of like we adopted this from like pyramid schemes almost right

like if you don't get in now brother and so I you know I think that one of the reasons why people dislike this technology is because of that but it's also part of like the paranoia, right? If you talk about these tools to try to get people to think that they're so powerful by saying they're going to destroy humanity, they're going to take away your job, they're going, you know, they can reliably automate all human processes. There's a company, I believe it's called Superhuman that talks about basically creating bot farms out of thin air,

right, to market anything you want to do. And the tagline of the company is, never pay a human again. When you say things like that, people will eventually listen. It's not really going to land well. Or it does land. And I think that's what's happening is it's landed. And people are like, oh, crap. And they're thinking, what is a human for next year?

Why am I paying to send my kid to college? Does my kid know how to read? Does that matter? You know, all these big things. And the idea that people could be left behind, it feels really salient now, in part, because so many people have been left behind already, right? Like, the path of predictable progress, to quote, the economist Kyla Scanlan for Gen Z, it doesn't feel real anymore.

And so all these fears are magnified. There's a software engineer Fernando Barretti who just wrote one of the bleakest pieces I've read about AI replacing us, where he argues that, quote, if there is a permanent underclass, you won't escape it by owning property or shares in anthropic or open AI or. guns or anything else, and neither will the billionaires. You, me, Sam Altman, Dario, everyone who is made of flesh and blood will be disempowered and replaced by machines. Basically, his idea is, in a world where AI does all work at human level or better, the permanent

overclass of rich people won't just be useless to the functioning of the economy in the state. They could be an obstacle to the state getting what it wants. For example, if the state goes to war, it will need rich people's planes and factories or whatnot, and then it will just take them. And then eventually there will be people running the state, but there will be so many AI machines that the states that have more AI will be advantaged over states that are still using people, humans in the loop, as you mentioned in the piece. And eventually humans will just be like, either the machines will let us hang around, or they just have no use for us and we'll just be like, basically pets.

How far-fetched do you think that idea is at this point? I think it is a little far-fetched, you know, it's tough to say. I think so much, I think it's better to think about power dynamics in all of this. Something that I really appreciate, talked to Corey Doctor-O on my podcast a couple weeks ago. He's got a new book about how he uses you. We talked about it. Yeah, he's great. Yeah.

And, but, you know, he uses this term by this, but this writer Dan Davies called Accountability Sinks, which is basically when some, like, it has a lot to do with the idea of that human in the loop, right? If you put the human in the loop, if you fire a lot of people, but then you put the human in the loop, and there's kind of one human who's like in charge of overseeing

what these models are doing, but they're doing it on this scale where there's so much stuff happening that it's very difficult for, you know, a human to wrap their head around every single thing and evaluate it. When something goes wrong, the AI doesn't get blamed, right?

When the software decides to, you know, target civilians in a battlefield, right? A human gets blamed for that. The person who was hired, you know, arguably for less than, you know, they were before because the AI is doing so much of this. They're the accountability sync, right? They're the person who has to deal with this. And so it becomes this way to, to, to, like offload their responsibility and the blame and the power.

Do I, in terms of the, the dystopian part of, you know, all the job loss and, and, and, and,

and, and, and, and, and, and, I do think the part of that that I agree with is the idea that this is going to come for the bosses for the,

the, the, the, the, you know, like, Silicon Valley is, is sort of first on the chopping block in terms of a

lot of these jobs. They have found that like, oh, man, like this isn't, you know, automating the bus drivers yet. This is actually automating the guys, us, like the people who make the code, right? And things like that. And so I think the same thing could be true of a lot of bosses, a lot of, or a lot of, you know, C-suite-level type people, right? I mean, what is a CEO when you think about it? It is somebody who evaluates a ton of information and makes a big decision. right? They're not actually like doing the nuts and bolts stuff. They're they're supposed to

process everything and react based off of all the amount of data they can ingest. That's really what a lot of these tools do actually really well, is they evaluate a ton of information and say this is what you should do. So the CEO is actually, you could imagine that person being on the chopping block, right? So I think that I'm not sure where this is all going to play out with the job loss stuff. I think that we may be, getting ahead of ourselves in worrying about a lot of this, in part because there are so many people with power inside of organizations who are so excited about efficiency gains and replacing workers and also telling people that they've replaced workers because that's really helpful for your stock prices and things like that. But I think we're still in like, you know,

the first inning of all of this and to predict where it's going. I think all we can say is that it's probably going to be unpredictable. This episode is sponsored by BetterHelp. Not everyone experiences summer as an endless parade of hot dogs, vacations, and pool parties. For some people, life's woes don't abate when the temperature climbs and the days get longer. If you've got the summertime blues or are simply feeling overwhelmed, BetterHelp is there when you need it. With over 30,000 therapists, BetterHelp is the world's largest online therapy platform, having served over 6 million people globally. And it works, with an average rating of 4.9 out of 5 for a live session based on over 1.7 million client reviews.

Better Help therapists work according to a strict code of conduct and are fully licensed in the U.S. BetterHelp does the initial matching work for you so you can focus on your therapy goals. A short questionnaire helps identify your needs and preferences. If you aren't happy with your match, switch to a different therapist at any time from their tailored wrecks. We love therapy. We're a couple of therapy boys here. You know, when everyone always needs to talk to someone, especially at times like these. And, you know, your friends and family are sick of hearing from you.

So go talk to a therapist. You got to talk to a therapist. You just got to. You just got to. There's an endorsement from John Lovett. You don't have to say yes to everything this summer. Find support and therapy. Sign up and get 10% off at betterhelp.com slash offline. That's betterhelp.com slash offline.

In moments like these, it's easy to feel overwhelmed and even easier to feel powerless. But we are neither. I'm Stacey Abrams, and on my podcast, Assembly Required, I take on each executive action, legislative battle, and breaking news moment by asking three questions. questions. What's really happening? What can we do about it? And how do we keep going together? This is a space for clarity, strategy, and hope rooted in action, not denial. New episodes of assembly required, drop Tuesdays. Tune in wherever you get your podcast and on YouTube. Speaking of dystopia and bosses who unfortunately may not get replaced, I want to check in

with your friend and mine, Elon Musk. History's first trillionaire, briefly. I don't know if he's back now, but he did lose some of his value, so he slipped out of trillionaire status this week. Your June piece, the myth of SpaceX, landed right around the IPO, and your argument is that SpaceX's insane valuation, launching its IPO at a valuation of $1.7 trillion, despite only $18 billion in revenue last year,

is, quote, built on a story that descends reality. How's that? SpaceX is a lot of things. SpaceX is a real company that launches rockets that is impressive, that employs human beings, that provides a lot of satellites that are geopolitically valuable and also gives internet to remote places and airplanes and all this stuff. So it's a very legitimate company. SpaceX is also a meme, right?

It is part of what the authors Quinn Slobidian and Ben Tarnoff in their book, Muskism, financial fabulism, which is Musk's ability to sell this story about what he is going to do, about the potential of SpaceX, about asteroid farming, about, you know, interstellar space travel, us being on Mars permanently, all of this stuff that has not happened, that really has no predictable pathway to to happen. But it is this meme of expected value, of expected genius, of expected innovation. And SpaceX is also a financial instrument. And SpaceX is also a financial instrument. right? SpaceX is some is a vehicle that Musk used to take his AI company and roll it into SpaceX in order to produce a massive valuation, a valuation that has very little to do with the rockets and the

things that SpaceX actually builds and and produces. Instead, it's basically valued as an AI company because that's what's hot right now. That's what, you know, can get you that type of valuation. So SpaceX is all of these things, but when you put them together, it's basically the end stage. And I don't mean not saying it's over, but it's like it is the sort of teleological endpoint of Musk's project, his whole thing, everything that he has been working towards. And as such, the market has valued it in an unprecedented way, given him this unprecedented wealth. When he is losing and gaining more than the net worth of. Bill Gates in a given trading day. So crazy. And you think about what he's already done with that money, funneling it into the Trump campaign, using it to buy a social media platform that he can then basically change the algorithms of and use to promote his own values or have the news behave, you know, according to the physics of Musk's own ideology. When you put all that together, it's all without precedent. It is this chaos agent,

lodged in the entire global economy, but also our politics, also our culture. And I was comparing it to the old Matt Taibi line of Goldman Sachs as the giant vampire squid, right? It's this predatory instrument that basically takes and takes and fixes the game in its own right. He wrote this famously after the financial crisis. And I reread that article. And I reread that article. on the day of the SpaceX IPO and I was like man vampire squid is quaint for what Elon Musk is because Elon Musk if you look at this that what's happened during this before the SpaceX IPO SpaceX was like essentially you know bullying is probably too strong a word but making the like Nasdaq 100 like list this in its on its terms right it allows for for this like for retail investors to go in

kind of stripping down some of the usual protections that are that are offered there. You know, they set all the terms of the IPO on their terms. Usually, like, the big banks, like J.P. Morgan and Goldman Sachs actually, you know, dictate, dictate the terms of what they're going to sell there. But not in the case of SpaceX, not in the case of Musk, because of all of that power. And so I basically said, I was like, I think Elon Musk has either won capitalism or just broken it, like, you know, used the game genie cheat code and just sort of, of like fried the gaming console on capitalism. And he's like the seven-headed hydra that sits at the end of finance, you know,

the final boss. And I think you said like he is too big to fail now. You know, I, I, it feels that way. I, I, since I wrote that article. So embedded in government and I mean, it is, I think you could unwind him from everything, but it's, you know, it's, I think his goal is to make that impossible. Yeah. Since, since I wrote that article.

it's not just that SpaceX's valuation has dropped, that he's not a trillionaire anymore, he's still so wealthy that that power is huge. But something that has happened with him in the last couple of days has been these claims from places like the Journal of the Lancet that Musk is responsible for upwards of millions of deaths of people in Africa and other places as a result of putting USAID into the woodchip and Musk has been just vehemently on X defending this kind of strisanding. Fighting with Rokana, threatening to sue Rokana over this. Yeah.

And to me, the fact that he, you know, he's dismissed this before, but the fact that this is like a real sticking point for him right now, that he is really concerned about this, it strikes me as a person who may be realizing either subconsciously or not, that the tides could change politically, you know, like not just with the midterms, but, you know, potentially in 2028. And the stakes may be real. Like, one thing that Musk has been able to do throughout his entire career is play all the different sides, right? Whoever's in power, he's been able to cozy up to the government, get lucrative contracts, get, you know, lucrative

investment, which he then funnels into the businesses and keeps, you know, all the spoils. But he's been able to do that. That has been the project of Elon Musk. It's been part of how he's been so successful. And I think what he's looking at right now is seeing that one of those pipelines could be cut off from an entire party, right? I mean, it's not only that like Trump could and has gotten bored with him from time to time, but it's this idea that like there could be, you know, like a truth and reconciliation style, you know, like commission on like what happened here. Like what really happened when you guys, you know, walked into the federal government and had, you know, 20-year-olds accessing all this government data and cutting all these programs that saved lives? And it's not just that he may face consequences for that. And, you know, maybe he will, maybe he won't. But it's that, like, he becomes radioactive to one of the, you know, the two parties in power.

That's a real vulnerability for him. That actually genuinely affects his status in the world. It is also, I mean, your argument in the piece is that so much of the value of SpaceX is built on perception and a story that Elon Musk is adept at selling. The flip side of that is, if that perception turns negative and the story unravels in the mind of the public, then there are real risks to the actual business. I think so. I mean, you've seen it already during the Doge era with Tesla, right? Yeah. With sales flagging and also with, you know, people having to put stickers on their cars.

It was like, I bought this before I knew about Elon or whatever. But it could go so, so much further than that, right? I think there's a lot of people who are able to either support him because Trump is still, the president, is still ascendant, is still, you know, has. a lot of power and therefore there's a lot of people in the Republican Party who are willing to just go along with this. But I mean, if the idea of what Trumpism is just gets resoundly defeated over the next two and a half years and everyone has to, you know, the Republican Party do that process of being like, I wasn't for that. Who? What are you talking about? If that happens,

I think you could easily see some of those people who've just completely sold themselves to Trump turn completely on this guy, right? Like we were all kind of under the spell of this dude we didn't have a choice and then yeah, this unelected man born in South Africa came in and destroyed our government

and we had nothing to do with it. Like that makes him genuinely radioactive, you know, and I don't think there's enough Musk fanboys on X to to, you know, keep Tesla or SpaceX or whatever afloat if that's the case. One Elon detail from this week that connects back to our conversation on agency, he sat down for a podcast with Peter Diamandis and argued that pretty soon money itself won't really matter because AI and robots will generate so much abundance that human work and maybe

wealth itself becomes beside the point. Do you think he believes that? Is it a sales pitch? and if it is a sales pitch, what is that actually doing for him? I was thinking about this the other day, and the person that came to mind was Alex Jones.

Because I've been writing and reporting on Alex Jones for like a decade, and obviously the huge question is, like, what does he really believe? Does he believe all this stuff that he says? And I think that the answer that I've gotten to is, A, it sort of doesn't matter. But B, it's like a hybrid of all of this.

Like, he believes certain, parts of it and other parts of it are just completely useful, right? And I think that like what Musk, what Alex Jones, with Donald Trump, what they all have in common is this, like their superpower is their shamelessness, right? They can take the things that they believe and use them to create the myth, but also take things that they absolutely do not believe and force them down people's throats in the same way because it's convenient for them. And so I don't know what he believes. His project has always been in a way that I think he's he's a genuine believer that like man and machine need to be fused in some way, right? But only certain men because he, you know,

he's really interested in protecting a certain demographic of people on the on the earth and advancing the light of their consciousness into the stars. But what I think about that is it's just, Again, it's a bad sales pitch, actually. It sounds to him like it's cool, but it's like what he's really doing is just talking about like the people in the chairs from Wally who are, you know, slackjawed and like roaming around nowhere, right, getting fed through tubes. And it's like, I don't know, man. That's like not really empowering. Again, agency crisis. Wasn't really the point of that movie.

Last thing. We always joke about how you never come on for the happy and hopeful conversation. But you do actually point to something hopeful in the piece about agency. You know that the backlash to AI is showing up in the physical world. People going to town halls, packing city council meetings to fight data centers in their own backyards. Is there a version of the next few years where we genuinely claw some of this agency back? One thing I've wondered, and I've experienced this myself too, is that our feeds become

filled with more AI slop. It might be, it's like breaking my addiction a little bit on the scrolling because I'm like, it used to at least be the illusion of connection with other human beings and you'd see their reaction. And now, especially Instagram, I noticed this too, if it's just, I'm like, is this a real, this isn't even a real picture. This isn't a real news thing. Like, why do I, this is, this isn't enjoyable at all. And I do wonder if it is going to maybe, maybe an upside here is going to push a lot of people offline back into the physical world, but maybe that's just too poly-inish. But what do you think?

My big prediction for this year, which hasn't come true yet, is, but maybe it is in some ways that I can't really see, is that, and some of it is the AI slop driven stuff, but some of it too is just the, it's the maturity of the internet in this way. this researcher Aiden Walker wrote this or actually it was in a TikTok but has this theory it's called like the cuck theory of the internet right

which is just that like when you're watching TikTok or an Instagram reel of like you know an AI cute animal or whatever or just something that's like generated in that way like what you are actually doing is you are sitting in the in the cuck chair

and on the bed is an algorithm that is trained off of all of this artificial intelligence data at the same time and your data and everything and this AI program which is understands intuitively the algorithm and they are you know doing whatever they're doing and you're just this like consenting observer you know you're just kind of like oh that's neat like the algorithm showed that to me because it knows what I like and this thing also knows what I like because it's fed off of everything that I've ever done and I'm just going to sit here and you know let a little drool come out of the corner of my mouth as I as I you know

enjoy whatever this is, is a fleeting pleasure. And I think that, you know, that's a funny way to describe it. But I think there is a feeling that a lot of people are getting, whether it's fakery or not, or whether it's just, man, this thing is just so absolutely attuned to me. It is like, like, to feel constantly like you are being played like a fiddle all the time, you know, that I think it is, it's a bit soul-crushing. And I think that when you take that and then you also, you know, add into like the element of surveillance and our phones and this feeling that I think a lot of young people have of like it's stressful to go out in the world and like take risks and take chances and go on dates without being filmed by, you know, Project Veritas or whatever, you know, it's like there's all these feelings that it's tough to go out and exist in this way. artists are feeling this way. Like, I don't want everyone filming my concert a million different ways

and sort of like I can't connect with anyone in the audience because they're capturing something for later. So you have bars that are starting to be like, let's check the phones at the door. Concerts that are like, please, like Phoebe Bridgers just said for this tour, hey, I'm not going to let people film. Sorry, like enjoy it. Be present. If you don't want to, you can stay home. And there's some backlash to that. But I think that it is this feeling in the world. was that the phones are going to become really uncool or this avatar of this being cucked by technology. And I do think that AI could hasten that, right?

And this feeling, people going out in their communities and protesting data centers. There's a lot of weird reasons that people protest data centers from all over the political spectrum. Some of them are kind of strange. Some of them are really real and great. But what people have noticed is there's been a lot of data center projects that have been canceled in the last year. And that is actual political activism that has, you know, that has results.

It works. And it's being in your community with other people, standing up for something. And I think that that is a real bomb. You know, we talked last time, I think, about like the ICE protests in Minneapolis. and the way that that is like, that is taking something that people see through their screens and then dropping the screens and then being in community with people. And it got results as well. And I think if we continue to see people of any political persuasion, of any way, seeing that being in the world gets a result in a way that a like or retweet or a post or whatever just can't give you,

I think that that's powerful. That's like the kind of reinforcement that actually builds societal change. That's a great place to end, making sure that we don't allow technology to put humanity in the cut chair. I think that's great. Charlie, thank you as always for joining offline. This was fun. Take care. Offline with John Favro is a crooked media production. Our show is produced by Austin Fisher, Emma Ilich-Frank, and Anisha Bannergy. Our team includes Dilan Vilanueva, Mia Kellman, Charlotte Landis, Eric Schute, Rachel Gieski, and Will Jones with support from Adrian Hill and Matt DeGrope. Our staff is proudly unionized with the Writers Guild of America East. In moments like these, it's easy to feel overwhelmed and even easier to feel powerless.

But we are neither. I'm Stacey Abrams, and on my podcast, Assembly Required, I take on each executive action, legislative battle, and breaking news moment by asking three questions. What's really happening? What can we do about it? And how do we keep going together? This is a space for clarity, strategy, and hope rooted in action, not denial. New episodes of assembly required drop Tuesdays. Tune in wherever you get your podcast and on YouTube.