from 404 Media
The narrator just like suddenly started saying, it's not just that you wanna make sure that you're not paying taxes. It's also that blah blah blah. And it was like using these formations that are very common in AI writing. I know this is a human being because I've been listening to this person for years, but she wrote her script with AI, and I can tell. And it's not that well, there you go.
Jason:So, okay. I went to This is it though. This went to write the article. I went to write the article and why I didn't like it. And I was like, it's not just this.
Jason:And I was like, I'm doing it. I'm doing it. My brain is
Dexter:doing it. Did you say you were gonna delve into the topic more?
Jason:Hello and welcome to the four zero four Media Podcast where we bring you unparalleled access to hidden worlds, both online and IRL. As a reminder, four zero four Media is a journalist founded podcast and needs your support. To subscribe, go to 404media.co. Subscribers get full access to all of our articles and our archives, as well as bonus segments on every podcast and early access to interview episodes such as this one. This week, we're bringing you a special episode, which is a live recording of an event we did at the LA Public Library about a month ago with our friends Dexter Thomas and Geeta Jackson.
Jason:So Sam, me, Dexter, and Geeta talking about how AI is destroying the Internet, changing how we live and work, and is also impacting Hollywood, books, games, every industry, basically. We had a really great time, and I think that it turned out quite well. So, thanks to the LA Public Library for having us. Here is our conversation at the LA Public Library. Thanks everyone for coming.
Jason:This is very cool. Thank you to the LA Central Library for having us. This is amazing. I'm Jason. This is Sam, Dieter and Dexter.
Jason:What's going on? Hey. I think before we start, we have a short video to kind of like frame the conversation. We've been covering AI and AI generated media and all of that sort of thing for a few years now. Things have really gotten a lot worse.
Jason:And so to start, this is like pulled from my Instagram feed, my Instagram algorithm directly from a few months ago and this is sort of like the type of information ecosystem that we're seeing out there. So I guess let's be horrified and then we can go from there.
Sam:That's the inside of Jason's mind and that will explain a lot.
Jason:Yeah. Do do we did we like that? I don't know. Again, that a few months ago. And I think that most of those were AI generated.
Jason:There are a few non AI generated things in there because I still do get some content that real humans are making. But again, that's from a few months ago that, the technology has gotten a lot better since then. I feel like for a while I wasn't getting tricked like at all because it was all very surreal and weird like that. But over the last few months, things have gotten a lot harder to discern. I don't know.
Jason:Do you all see Are your feeds like that?
Gita:My feed isn't nearly as deranged, but I have noticed, like, I like to watch, especially when I'm in the bathroom, just random ASMR videos on Instagram and that has been completely taken over by AI videos. Especially, love clacky keyboard noises, and that is like a really really popular AI generation platform. And I usually don't get tricked by them, but they are inescapable if you are looking for that topic.
Sam:Yeah. And you watch one and you're you're suddenly watching 400
Gita:in a Exactly. Yeah. It's been fun watching people who it's usually like, what if this keyboard was made out of honey? Like, what if this keyboard was made out of marshmallows?
Sam:Or the beds.
Gita:Or the beds. Oh, you like, what is what it's like to sleep on a water bed that's made out of the ocean. Yeah. That kind of thing. Like, this gets sucked into my my algorithm.
Gita:And it's interesting to watch like, there's a YouTuber Rosanna Pascano, I think, who did this project where she was like, I'm gonna take those viral like glass fruit ASMR videos and see if I can actually recreate something that gives you that same effect in real life. And often, you basically just can. You can make a goopy keyboard made of honey. It will take ruining a keyboard and a lot of honey, but like that is an ASMR effect that is possible to achieve. But I imagine just sitting there and like typing prompts into a computer is much much much faster.
Gita:It sucks.
Sam:We were just saying before we came out here that I am getting tricked more and more often because the tricks that you used to use to detect fake images are no longer relevant. You it's not like people can blink now, you can't detect like a deep fake by watching people's eyes and blinking. There's like this viral it's like Memphis Jukin video that everyone has seen at this point probably, but it's like a dance that's very repeatable. And I at first was like, oh my god. Look at this guy.
Sam:He's like so good at this. And now I'm realizing that like all of these videos are actually fake, so I'm seeing them over and over. But it took seeing the same video over and over for me to realize, this isn't actually real. That's how good it is. And I look at this stuff all day long, so I should know better.
Sam:But it's just getting so much easier to just scroll by and be like, oh, that's real, and you keep going.
Gita:Yeah. The superficial details that were often big giveaways have more or less disappeared. But if you know something really specific, it often makes it really obvious. I know what a bicycle looks like, right? And AI video just sometimes has difficulty with like a bicycle in motion, like where all the things are supposed to go.
Gita:And so if you're familiar with something deeply familiar or a cable I worry so much about cable management in my house that when I see like cables that don't make sense in an AI video, I'm like, oh, that wire is not going anywhere. But it takes a it's not like you can't just look for text, you can't look for cars in the background, you can't look for wires, electrical wires going on telephone poles anymore. A lot of these things have been smoothed out and it's could be scary because you'll also see malicious grifters creating wholesale AI influencers that are often stealing content, often taking content from black women and putting a white woman in it, often trying to generate incredibly like, propagantic discourse. And they that two face two camera format that influencers use, it's just so easy to replicate that now and have it just look like a human.
Jason:Yeah. It's very interesting. I follow the kind of business model of a lot of the type of content that we just saw, like how it's made, who's making it. And a lot of it is somewhat kind of like incidental. Like a lot of it is not, people who are necessarily trying to trick an audience.
Jason:There's like a monetary, like impetus for it where they're just trying to go viral. They're trying to collect money, from the ad impressions and that sort of thing. And so for a while, people were saying, you can do faceless AI, faceless Reels, faceless TikToks. That's what they were saying because basically like the LA Fire stuff, it's like there's no person on screen narrating that. It's just scenery, that sort of thing.
Jason:But that's changing like recently as Geeta was saying, it's like you can digitally clone yourself, you can digitally clone someone else, create an AI influencer and have them read the news or have them say whatever you want. I want to talk a little bit about the like topicality of it as because the Internet is really driven by like the news cycle. Like when there's a big mass shooting or a war or a new movie comes out like that sort of thing, there's like journalists write articles about it, YouTubers make videos about it. And what we've seen is there's been a rush of AI content, being made about any, like, world events basically. And so immediately after the LA fires, Dexter did like a really amazing YouTube video tracking, the types of like fire disinformation that was coming and also, the ICE raids like in LA and in Minneapolis, all over the country.
Jason:There's a lot of AI generated content around that. And I was curious, Dexter, if you could tell us a little bit about like what you found. I know there's like outdated references
Dexter:No, the thing is it's not. I mean, I was just thinking about this earlier because like you haven't talked about shrimp Jesus yet, which is the amount of restraint that you're showing now is amazing. Everybody familiar with Shrimp Jesus? Okay. Yes.
Dexter:Literally because actually when we talk about AI slop, like we should probably talk about that Jason and four zero four in general have had a really big hand in people actually understanding what the phrase AI slop is. I know that you don't necessarily think you created the term, but it's something that, right? Like you were
Jason:was one of the very few people reading fascinated by broke my brain with it by just like looking at it all day every day.
Gita:Every day in Slack, different pictures of Jesus Christ made out of shrimp or Jesus Christ being eaten by shrimp. Just like I I didn't wanna see these things, but you showed them to me and now I know about them.
Dexter:Jesus Christ made out of gold, Jesus Christ, being lifted out of the ground by a helicopter or something like this, all of this stuff. And and so just to show you the progression here, think, you know, late twenty twenty three early no. Late twenty twenty three, I think. No. Yeah.
Dexter:2024 also. We talked about this and when we talked about it, the whole time, just to show you, we were laughing the whole time because it was sort of like, isn't this funny? Our parents are on on Facebook and being fooled by this stuff. Right? Fast forward about six months, it was like six to eight months after that, I'm getting to the fires thing.
Dexter:I'm on Instagram and the fires, fires are still happening, but the fires are still happening here in LA. And I'm seeing this video of and I think the first big video that I saw is there's a video of like a firefighter carrying some baby bears to safety and there's like a fire behind it. And it all looks pretty real and it takes me a while to realize that it doesn't really make any logical sense. Like if you keep watching for example, there the there's another firefighter that's like picking up a raccoon, but he's like pushing the raccoon into the smoke like instead of like saving it. But if you're not really really truly paying attention, you don't quite pick up what's going on.
Dexter:But anyway, by this point everything looks really real. It looks really real and in the comments, so we can talk about the comments, you see people saying, Oh my gosh, I'm so glad that people are saving these animals. Like everybody's worried about the people but we really need to worry about these animals. I mean, disagree with we should save the animals, they're all cute, right? But then the other interesting thing here was that occasionally somebody also would notice, hey, this is AI.
Dexter:This isn't real. And the responses from the original person. So you have person a saying, hey, I'm glad this look at these brave firefighters are saving the raccoons, are saving the baby bears, are saving the baby deers. And then person B says, hey, this isn't real. Person A says, how do you know?
Dexter:Or they say, I'm sure this isn't the same person, but I'm sure this is happening somewhere. This is somebody trying to raise awareness. So why are you getting upset over these little details when we should focus on the truth? The truth is that this is happening somewhere and you know, you see similar things with you know, ice raids or whatever. It's like you know, you see all these images of people angrily storming buildings and all all these things like that and it's like, yo, this this actually isn't happening, man.
Dexter:Like, I'm outside. Like, I'm physically in front of the building that you're saying they're burning down right now. I'm like, the building is still there and I'm seeing these comments coming in. It's like as I'm physically outside watching federal agents shooting tear gas at people and other assorted things. And people are saying, well, I'm sure it's like this.
Dexter:It's just they're not capturing the actual angle but this is a dramatization or whatever, but this is what's actually happening. This is the truth. And that's where you start to realize, oh, okay. This isn't really about fact checking anymore. This about what feels real.
Gita:It's what people would really like to see and want to believe. Another genre of ICE raid video I saw repeatedly was a pastor standing up to an ICE agent. And like that is something, it feeds something inside you that you want to see. You want to see resistance if you are something that someone that's sympathetic to immigrants, I hope we all are. And you you want to see someone doing something that you wish you could do even, which is stand up in the face of danger.
Gita:And the fact that it's not real, there are a lot of people kinda just don't care about that. They just want the warm fuzzy, the dopamine, the serotonin you get from seeing something that affirms your beliefs and makes you feel like some good is being done in the world. And they'll say the same thing, that's the same thing. Yeah. I'm sure someone out there is doing it.
Gita:I'm like, it's a way of abdicating responsibility away from yourself also because you could be the person that's doing that. You could go to a protest and stand up to a nice agent. That's dangerous though. So you'd like to watch a video of someone doing that instead.
Dexter:And so I think a lot of this is it's a there's a system here which we have to remember. I mean, you were talking about like the incentives and like the fact that you can make money off of this stuff. Right? And so that's what a really important point that's brought up because I think when we think about AI, we think, oh, this is a computer, this is a robot that's making it, but there are actually generally people behind the machine that are making these things. And so, is everybody familiar with a Meta?
Dexter:Do we know what a Meta is? Not Meta the company. Not the company. Meta okay. Meta the so the Meta game.
Dexter:Everybody know what a Meta game is? Okay. So meta meta game is the game that's on top of the game. So for example, Tic Tac Toe. Tic Tac Toe is not there's a way to win the game.
Dexter:There's a way to almost guarantee that you win the game which is to go first. Everybody knows this, right? If you go second, you're probably gonna lose. Just go first. That's how it works.
Dexter:Another things like say if you think of like, Street Fighter, games like that, people realize that, oh, alright. It turns out that it kinda matters what character you pick. You pick this character, there's a best character or pick this character, stand down at the corner, just throw fireballs the whole game and you'll probably win. Like you're probably not gonna get hit, it's really easy to win. So that's a game on top of the game.
Dexter:People realize that there are ways to quote unquote hack the system. Right? So there is a meta game, there is a meta and so when we say meta, that's like the best way to win the game. So think of social media as a game or think of capitalism as a game. The best way to win the game of capitalism or the best way to win the game of social media capitalism which is to say, to get the most views.
Dexter:What you wanna do is you wanna do something that feels emotionally impactful. You just want a reaction. It doesn't have to be true, it doesn't have to be insightful. It doesn't even need to be necessarily like mean or wrong or like mean spirited. For example, seeing a video of somebody standing up to an ICE agent.
Dexter:Hopefully, everybody in the room would think that at the very least you have a first amendment right to go up and say, hey, I don't think what you're doing is a good thing. That's legal, right? And there's probably some people who in the room would say, yeah, that's a good thing that they're doing. If you can make that if you can make a video of that moment feel real, you don't have to care as a person creating it whether or not that is actually happening. So the person who is making these videos of the the fires in LA, these baby bears being take being rescued from the fires and stuff like that, I talk to them.
Dexter:Hit them up, I DM them, they talked to me. They're Russian. They don't live here. It doesn't matter what their background is. What matters is like they have no stake in this other than are people going to feel emotionally attached to this?
Dexter:Yes or no. As the fires started to fade from the news, they switched to other stuff. There was a rumored ceasefire going to be happening and so they posted something about Gaza, that video flopped. They moved on to something else. I think if you look at their video now, posting cute animals because they're starting to think, okay, well maybe people don't care about fires or like genocide, maybe they care about cute bears.
Dexter:Let me go back to the cute bears.
Jason:It's all I think. Think that's the really important point is that everything you just said is totally true but the thing that AI is breaking is the scale and speed and cheapness that this can all happen at. And we're talking right now, we're talking about swap on like Instagram and whatever, but it's happening to every sector of the economy. It's happening to everywhere on the internet. It's like SEO used to be a thing, search engine optimization where, you know, you wrote If you wrote recipes, you'd have to write an entire like back story about why you're writing the recipe so that you can rank higher in Google and so on and so forth.
Jason:And human beings used to write that content and they started doing this like formulaic, here is why I make chocolate chip cookies. This is my why for that. Like, here's a 3,000 word article about these chocolate chip cookies. And that is all there's now this thing called AI engine optimization where that same thing is happening, but the content that's being created is being created so that OpenAI scans it, that Anthropic scans it and that ends up in these AI engines. And so like the point of this panel is to talk not just about social media, but what's happening to Hollywood, what's happening to books, what's happening to, like influencer culture, what's happening to like the porn industry, what's happening to everything around us because what's happening is the new meta is spam.
Jason:It's just like spam across everything.
Dexter:Yeah.
Jason:And it's not just on social media, it's like the entire internet, the entire economy. And the reason that it's spam is because for a human being to make anything or to do anything, it takes time, care, effort, money, resources, like all that sort of thing. But to make AI generated content, the entire purpose of it is efficiency and maximalizing what you can do. And so you can basically just spam out like, oh, this type of content is not working. Well, it only took me ten seconds to make it.
Jason:I will try something else. And when that other thing starts working, you just do more of it, more of it, more of it, more of it. And it becomes a lot harder for like human creators, human authors, human influencers to compete with that.
Sam:Yeah. I want to go back to what you said about the recipe thing because I think that is an important point to make. I think we're in a weird position right now where all the AI stuff can feel very tricky and fresh because it is newish. It's in the last couple of years. I read about the adult industry and the porn industry quite a bit, and something that that industry in particular is having a real crisis about Like, other people are having a crisis about it.
Sam:The people who are actually in it are not actually that worried, but it's like our AI influencers, I'm sure everyone has seen If you're on Instagram at all, you've seen AI influencers who are like doing some kind of OnlyFans grift or like they're pretending to be like hot girls to then get the comments from men who are like, you look so beautiful today. And it's like, that's not a real person. But it doesn't it doesn't actually matter to the people who are commenting. It seems like they know that it's AI, but they just want to participate in this ecosystem and feel seen and feel like they are, like, part of something with all these other guys in the comment section. But something that I've started seeing with the AI babes, as we call them, is they they've started needing plot lines of their own.
Sam:They need to pretend to have lives and interesting like It's like with the fruit dramas or those kinds of things. It's like they actually need to have a human emotional pull that goes along with just like the hot girls cause that's not You can Google, you know, like hot girls and hot girls come up. But like, you actually need to feel an emotional investment
Jason:in Backstory.
Sam:Right. They need a backstory.
Dexter:They need lower.
Sam:Friends. It's like these hot These babes have friends and they're AI friends and like they talk to each other on their other accounts. So then if you're following this, you're kind of following this like micro drama of the AI babes, but I think we will reach a point, and this is why the porn industry is not that worried about it because they know this very intrinsically, where they're what they're building that on is human creation, human content, and IP, and also just like experience. So eventually, they will need to do the equivalent of like the recipe thing where you need to have an emotional pull, not just like, I like looking at this eye candy. It's like, I actually wanna invest my time into this where the humans who are in this industry aren't that worried because they know that the long game is having loyal fans, having people who care about what's going on like in your life even though that is like a parasocial connection, caring, you know, what you had for breakfast today, what your boyfriend's doing.
Sam:Like, they they feel a connection with these people that they crave that they're maybe not getting in real life or that they want supplementary to their real life.
Jason:Yeah.
Gita:And I
Sam:think that is like that gives me hope honestly. These conversations get really depressing really fast, but like, that gives me hope in that like there's gonna be some kind of moment where AI can't actually keep doing that well and it will have to plateau because it's drawing on us, it's drawing on human experience.
Jason:I wanna, jump on that to ask you a question. Year or two ago, I, went to the Chinese theater over here because TCL, which is a Chinese like television manufacturer, was launching a streaming channel that was going be fully AI generated and they were showing these movie shorts. And there's this dream in Hollywood that, you know, you will be able to use AI to make more movies, editing process, like script writing, all of this. And to some degree like this is happening where like human beings are being laid off and the ones that are left are being asked to use AI. But some are just saying like, well, maybe we can just like AI generate the whole thing.
Jason:And they launched these movies and they were very terrible and they were like, we're going to do this. This is the future of Hollywood and, you know, there's some that are like not so bad, but the ones that they were doing were like extremely bad.
Sam:Jason was like, how mean can I be in this story? And we were, mean.
Jason:Yeah. Like, these were getting, like, 30 Like, 30 views on YouTube. Like, nothing. Like, terrible. And backstage in the backroom, if you will, we were talking about backrooms.
Jason:Yeah. Geeta, you made the point that like things that are like very practical and very human are really breaking through now.
Gita:Yeah. So if you don't know the background of Backrooms, it started not just as a viral YouTube, series, but way way way back in the day of creepypasta, copy and pasted ghost stories essentially that people would send to each other. This was an image that was posted to 4chan of a furniture store being disassembled and just a picture of a showroom essentially without any furniture in it, which just gives you a creepy, uncanny feeling. So from there, that turned into several different independent media projects, and then also has turned into this movie directed by a 20 year old, which is kind of mind blowing. Had a budget of $10,000,000 and this opening weekend it made, I think 80 or $90,000,000.
Gita:It is I saw it yesterday. It's a great movie and I read an interview with Kane Parsons where he said that 90% of the effects in the film were practical. And you can really tell when you watch the movie that they're in a real space surrounded by real objects. Disney, is a corporation that used to have a deal with OpenAI, has expressed a lot of interest in adding AI into their workflows. There was also Amazon recently announced a show, an animation show with that was going to be their foray into AI animation and one of the principal animators on it, this man named Jorge Gutierrez was harassed so badly for being a part of that project that he has stepped off.
Gita:But he has also said, AI is already in the animation pipeline, it's inevitable. But what I feel is, I look at a movie like Backrooms and how much money it made and I also look at like, I Love Boosters where you can tell in I Love Boosters that a lot of the effects are practical and also when it would necessitate VFX in some way. They they do stuff like claymation, stuff that's really physical and obviously looks like isn't the point isn't realism. It looks fake, but it looks it's like a movie. And audiences are really responding to the evidence of a human being, essentially.
Gita:I think I've also noticed this just with my website Aftermath, it isn't just that people want to sign up for a website because they think our writing is good. They like knowing for a fact that human beings wrote it. And it gives them an ease like a sense of ease to know that we are not using AI, we will never use AI, and the words that you read on your page, they come from real human beings.
Sam:Jason, do you wanna talk about on that note why AI is breaking your brain? Because it is like the exhaustion of like, you don't you can let your guard down. You can think, a human wrote this so I don't have to like fact check it or think about it. And you wrote recently about just like AI vigilance breaking your brain.
Jason:Yeah. So in the lobby, there's zine that I made two days ago, in my garage.
Sam:A human made that.
Gita:Can
Jason:It's tell very lo fi. It's not the best There's the cool
Gita:lesson in the back if you need any reason to pick it up. There's the cool lesson there.
Jason:But basically inside it is an article I wrote a few weeks ago that I was walking my dog in Venice, I live in Venice, and I was listening to a podcast that I had been listening to for like years and the woman narrator just like suddenly started saying like, it's not just that you want to make sure that you're not paying taxes, it's also that blah, blah, blah. And it was like using these formations that are very common in AI writing. It's not just this, it's that. And she did it like 14 times in a row. And I was like, this is a human, I know this is a human being because I've been listening to this person for years, but she wrote her script with AI and I can tell.
Jason:Like I wasn't even trying to pick that up, but like my brain was like, Oh, like I can tell that she did that. And it's not that, well, there you go. Okay. This is it though. I This is went precisely to write the article and why I didn't like it.
Jason:And I was like, it's not just that. And I was like, I'm doing it. I'm doing it. My brain is doing it.
Dexter:Did you say you were going to delve into the topic more?
Jason:Did. I'm I'm just That's not in my vocabulary really, but that's the thing. It's like these machines are trained on real human writing. They have these patterns, it's pattern recognition. They start doing it and I don't really want to be like the AI police, the AI cops being like, did like my colleague write this using AI?
Jason:Sam, did you send that email? No, but I get all these emails from like readers and things like that. I was like, or PR people like, please write about my company. I was like, this is AI generated. Like you didn't put any effort into this.
Jason:But what I noticed is that my brain is doing these like constant machinations like, is this real? Is this AI? Is this real? Is this AI? Do I care if it's real?
Jason:Do I care if it's AI? Is it low stakes? Like some of the stuff that I showed you up there, I don't really like any of it, but some of it now that I'm getting, I'm like, that's not so bad. Is that okay that I like it? And I feel like everyone is doing this to some degree when they go on the internet.
Jason:And I think it's pretty maddening, honestly. And I guess, think that it is undeniable that artificial intelligence is changing the world at the moment. It's like, there's so much money being put into it. Like people are losing their jobs because of it. You are encountering it.
Jason:Like, I think we can resist it. I think that you can be very human and sort of try to cut through the noise in that way. But like companies are firing tens of thousands of people at once because they're like, well, we replaced you with a robot. And so I don't think it's something that's just going to like suddenly go away, but this article is basically about like, how do we deal with this? Because we are in a situation now where everything we encounter on the internet, but also in real life, like I see menus, like I'm like, this is AI generated.
Gita:A lot of Spencer Pratt's campaign materials.
Sam:Can we talk about Spence?
Gita:Oh my god. Please.
Jason:Let's talk about Spence. Please save me. Let's talk about
Gita:anyone trusts him when he was on the hills. We have like documented evidence of what he's like as a human being anyway.
Sam:Do you wanna get into Spencer? Because you this has also been ringing your brain.
Jason:Yeah. So I mean, obviously, Spencer Pratt running for mayor of LA. I don't know if everyone here is like familiar with the type of campaign that he is running, but the type of campaign that he is running is one that is predicated on going viral repeatedly on the internet, using AI generated videos where he's like wearing a Batman outfit and he's like attacking Karen Bass, like at a a little Yeah. Naughty And you know, the thing is that this, he keeps going viral. He's getting tons of attention.
Jason:It's unclear whether any of this attention is like from people in Los Angeles or whether it's like Silicon Valley AI boosters and Elon Musk on Twitter and so on and so forth. But he's gotten like real buzz by doing this and it feels like one, the new way of kind of like campaigning, but then two, a lot of these videos he's actually not making, it's like his fans are making them. He's picking up the ones that he thinks are popular or like, I don't know, make him look like he owns the libs the most or something. And then like he's posting those. And so it's almost this, situation where he's just like grabbing stuff off the internet and using it.
Jason:And I don't really know what to say beyond that other than it feels very bleak and dire and I don't really It's like so decentralized at this point.
Gita:It reminds me a lot of an article I read in the New York Times a couple months ago that I responded to on Aftermath about AI romance novels. And there were a lot of There were, like, three authors in there that said they use AI in some part of the process of writing romance novels. But then, like, I did a little bit of digging into the first author that was quoted most extensively who coincidentally sells a course to teach you how to write a romance novel with AI. And she essentially writes just as many books as possible churning them out formulaically and they all sell maybe 30 copies each. And then you look on Goodreads and they all have like one or two stars maybe.
Gita:People don't actually like reading them. Mhmm. But it takes something that is like real, like something that's real. People like reading these books. They may give the people comfort, they make them happy.
Gita:And then it's reducing it just to its most cynical components, you know, campaigning as just I need to be the most popular. It doesn't really even matter what I'm saying or the how I am owning the libs. But if someone makes content for me and I look good, then I'm just going to recirculate that. And it feels I don't know. It makes me depressed because it feels like there's some kind of social contract breaking down there.
Gita:Like, I I want to I don't want to be in a community. I am working on a romance novel right now, and I don't wanna be in a community with authors that wanna take this shortcut. I like to write words because I love writing, and I love the craft, and sometimes it makes me wanna bang my head against the wall. But, like, that's, like, what makes writing good and fun. I've I've talked to programmers who say, I know that Claude Code can write a lot of basic programs faster than I can.
Gita:But I code because I like the craft of coding. Yeah. And I don't want to offload that to a machine. And here where it's about like advocating for an entire city, it just feels disgusting. Like, it gives me the like, hugely.
Gita:And I feel an intense disconnect with with people who find that appealing.
Dexter:Well, hold on. You're not just writing a romance novel. You signed a book deal. Right?
Gita:Yeah. Hope I I'm Tentative publication date next year, next to fall.
Dexter:Don't undersell it.
Gita:Come on. Yeah. No. Come on. Level Up For Romance is coming from Mira Books next fall.
Gita:Yeah. It's, I have always wanted to write a novel and to see these authors that are essentially saying that AI novels are inevitable. Eventually, everyone will use AI in part of the process. If it's idea generation which is like AI doesn't have any original ideas definitionally, and or or proofreading or something. Frequently, like it's like you don't even like writing.
Gita:You don't even like writing. Why do you wanna be an author if you wanna use something to think for you to create this product? But it's made me sort of confront this fact that, like, a lot of people actually do want to skip all of the hard parts and just have the success. It does sound really appealing when you put it that way. Right?
Gita:Like, is a a tool that appeals so directly to scammers. And I now just aesthetically associate these images and that kind of writing with someone who's trying to pull the wool over my eyes in some way.
Dexter:Well, I mean, okay. So this is man, there's a lot here. So that is that brings me back though to what we were talking about earlier though, which is it's the meta. We've realized that the meta is just volume.
Gita:Yeah.
Dexter:It's volume and right, know, what Wayne Gretzky, you miss a 100% of the shots that you don't take. Just take a million shots. One of them will go in. They don't have to be good shots. You just like hit the puck in this general direction which AI can do that for you.
Dexter:You can hit the puck in general direction for you. One of these things is gonna go in and that's it. I mean, it's meant to practice not to make a too big of a deal out of this but that model that you're talking about which is other people posting, making clips for you. That's the Andrew Tate model.
Gita:Yeah.
Dexter:That's the Andrew Tate model which is just to say that it's not really about having anything specific to say, it's about getting a bunch of other people to make clips for you and then you promote those clips and those will go viral and the person who's making the clips quote unquote, big scare quotes here, everybody wins. I think society loses to be very clear here. Everybody wins, right? But I think okay. To talk about another game thing though, not a good chess player at all.
Dexter:But one of the things that I hear from chess players is that they can recognize, like a really good chess player can recognize when somebody's setting up a particular move. Like they know what okay. I see what you're doing there. Right? So if I the thing we're talking about here where, you know, people want human connection and they want to feel human connection.
Dexter:I think if I'm to look at this from the perspective of an AI booster, right? Take a 100 steps back here. I think what is happening here is that yes, maybe some of us in here truly care about the craft of films. Maybe some of us in here truly care about movies. That Venn diagram isn't necessarily a circle though, right?
Dexter:Maybe some of you really really really care about filmmaking, maybe some of you really really really care about writing, maybe some people some people really care about food. There's a couple restaurants here which I will not name which I know the people who own them and I've talked to them and they they'll tell me and I talked to them when they were opening, they were saying man we make this really handcrafted food, it's all organic or whatever and I look at the menu and all the imagery is AI generated. Food is very important to them. Visual art doesn't matter, it's decoration. We've gotten to a point where like if you ask for organic food, you're a snob.
Dexter:That's weird. Really? Like seriously like and I'm kinda one of those people. Like if you tell me like I only eat organic tomatoes like bro, come on man.
Gita:Give me the chemicals.
Dexter:Seriously, like give me all the chemicals for real. But I think this is what's happening. I think that the move that's being made and I don't think there's like some group of AI people that are standing to make money. But if I was thinking from that perspective, it is this. Some people will resist in some directions but they won't resist for everything.
Dexter:Not everybody really cares about like I really care about music. You play AI music around me, we're probably not friends anymore.
Gita:Yeah.
Dexter:You know, like the images like I might give you a pass. Kraft macaroni and cheese, let's go. Like I don't care. Like AI that up. Like make it more artificial.
Dexter:I don't care. Right? But this this is what I'm talking about is that not everyone, not every individual is going to push back on everything and want human in everything. That is the move that's being set up. Now the problem so, okay.
Dexter:What's the counter move? Bro, I don't know. That's the problem. Is it sometimes you're playing chess you're like damn that's a good move. Not really sure what to do about that but I think it's good to know what move is being set up and I think that is the structure that's being put up is again that there is a the bet being made is that not every human being is going to demand humans in every aspect of their art.
Gita:Yeah. There is a video game that is very popular and very successful. It's an extraction shooter called Art Graders. And my cousin loves this game. When this game was first released, the developers were using AI voice acting for some of the they're called barks when characters say things off screen like, oh, a grenade, you know, that kind of thing.
Gita:And if you're, not really cognizant of like how video games are made, like you might think, oh, having AI barks is probably fine, know, an AI voice going, oh, you got me, is probably not that bad. But there's a whole industry of real human beings that do that for a living and they normally get paid for that. And like now that industry is beginning to shrink. And what was hilarious though is when art creators started making a lot of money, suddenly the development team to win back social goodwill from other gamers were like, we're gonna replace all those AI voice lines. We realized that actually human voice acting is just better.
Gita:And I find that really interesting They know that it can cut like good press and PR from websites like Polygon or Kotaku writing art creators developers to remove AI voice lines. But also, you decided that human labor was not worth it in the first place. So if you already knew that a) like human voice acting was better, why didn't you just hire a voice actor?
Jason:Well, I think that's a good question, but I think that's the thing is like the economics of all of this is like, that's why we try to focus so much on like, what is the economic incentive behind driving all of this? And I think we do use AI in different ways at four zero four Media, but not generative AI, not like that. It's just like Google search is AI, Google maps is AI. There's different Like Translation. In
Gita:animation, line stabilization is a tool that uses AI like predictions. And that is something that animators really, really love.
Jason:It's been, but like these different like translation is a heavy
Dexter:topic too, man.
Jason:Yeah.
Dexter:No, but like all It's
Jason:been like injected into like every part of like Adobe software, like just using the internet, using different types of machine learning and these sorts of things. Not all of it is large language models. Actually a lot of it is, a lot of the most useful stuff is not. But I think that each person sort of needs to decide like what is their line for like, what is your creative process going to be? Or like what types of AI are you going to use in your work?
Jason:And like how does one abstain from it or resist it if they want to? Like I did an article a year ago basically being like, the journalistic pivot to AI is not going to work because at like big media companies, they're like, we are going to have robots write our articles. And, you know, something that I think about a lot is like, takes so much time and effort to gain credibility and trust and it takes like zero seconds to ruin it. It's like some weirdo wrote a book recently and it's like half of the things in it were AI generated and the New York Times wrote about it and so on and so forth. And it's just like, I don't want to read that book anymore.
Jason:Like, I forget the name of it, but it's about how AI is changing everything. Half of the sources in it are not real And it's like, cool. Good.
Sam:You're not wrong. It is changing everything.
Jason:Yeah.
Sam:It for sure is. Yeah. And like we've we've made active decisions to It's like, so you have like hired people to do like, actually, like, translating YouTube videos when you were investigating a lot of this AI stuff. It's like, I need someone who's, like, a native speaker of this language because the auto generation that the AI is doing on YouTube is not good. Or, like, if we wanted to We did the zine last year that was about, ice action in LA specifically and flock and surveillance, and we hired translators to translate the zine into Spanish because the work that it would take to run that through, like, Google Translate, and then we would have to hire somebody anyway to check it because you can't trust the output anyway.
Jason:We don't run it through Google Translate and then just published it and then someone would be like, you screwed this up. Would be more I don't trust you to use AI, like whatever. But yeah, I mean, to your point, like a lot of the AI slop articles that I've written, the sort of the industry behind it, if you will, is in India. And so there's all these YouTube videos in Hindi about like, here's how to generate videos about the AI Jesus. About Shrimp Jesus and the LA wildfires and here's what Americans care about in all of this.
Jason:And it's like, I don't speak Hindi. I don't speak every language that these YouTube videos are happening in. And I guess in an ideal world, I would hire someone to watch all of those and tell me like what's going on here. But honestly, was like hundreds of hours of videos. And so what I did was I watched them using the auto translate feature, which I guess is AI.
Jason:And then I was like, this seems interesting. Let me now pay a human being to actually translate what they're saying. Like I was using it to sort of get the vibe of what was happening in any given video. And like to me, I felt like that was an, I mean, it was the YouTube built in like, translation service. Like I didn't feel like I was like sinning, if you will, like to do that.
Jason:But, and then we had that like extra layer and same, like I write a lot about FLoC, the automated license plate readers that Sam just mentioned. And it's like all of the things that are happening in that world, all the resistance to FLoC right now is happening at city councils and like around the country. And they're happening in these meetings that are like four hours long and I'm not watching like four hour long meetings from like 50 different cities every week. I'm opening up the transcription tool on YouTube, control F for flock. And then I'm watching like part of it and I'm like, Oh, this is the meeting where they talk about flock.
Jason:Because sometimes it's a four hour meeting. I thought they were gonna talk about flock and they don't at all. And I was like, cool. That's four hours of my life I'm not getting back.
Gita:Do you guys still use Otter for transcribing interviews? Otter, I'm horrible, horrible at transcribing interviews because I hate the sound of my own voice and I just don't wanna hear it. And I, Otter is very useful for just turning a really long audio file on my phone into something that I can go back into, control f for certain keywords, and then sort of double check the transcription to make sure it's a 100% accurate. And like, otherwise, I'd be stuck here not transcribing anything because I don't wanna hear my stupid voice.
Sam:Also, it's like these things are getting worse as they as time goes on because they're ingesting more and more of the same bad shit. And so, like and now Otter is also not to just sit here and dump on Otter, but they should fix this stuff. It's like they're ingest they're injecting more and more AI prompts. It's like every part of the screen now is like, do you want AI to do this? Do want AI to summarize this?
Sam:I'm like, no, I don't. I just want you to do the thing that I pay for, and then wanna actually do the work of looking through the transcript and looking through the interview that I just did and being like, this is important to me. I don't want you to tell me what's important.
Gita:Google started writing email responses, like prewriting email responses And for I like wanted to shoot my computer with a gun. I was so angry that day because it's like I wouldn't say that, number one. And two, like I can tell just from a context that you don't know what I'm talking about, what this email is about. And I had you have to go deep into its settings to turn off all like AI features and even then you still have a prompt for Gemini right at the top of the screen replacing the place where I usually would click to go get Google Calendar. So my mouse over accidentally all of the time.
Dexter:So this this is the issue though I think, is that we're all a very particular group of people here. They're on the stage. But also everybody who like the fact that you came here, you're a very particular group of people. I mean, okay. Actually, going back to like the the art thing, we were talking about like the act of creation.
Dexter:Right? How many of you would refuse to listen to music that was created with AI? Okay. How many of you would refuse to read a book that was created with AI? Okay.
Dexter:How about watch a movie that was created with AI? Okay. Some of these are the same, some of these are not. What what about use a computer program that was partially created with AI? Wow.
Dexter:How am I gonna know, but then less people, but then okay, what about the people who truly care about this? Who truly care about coding? What about something was translated with AI? Okay. See, now this is the thing and actually this is why I think what four zero four did was so important and I don't think if it really like flew over people's heads, but a lot of the ICE reporting was translated into Spanish by human beings.
Dexter:This is why I like the phrase slop so much, is because we're used to the idea that certain people get a degraded experience with something. If you can only read Spanish or Spanish is the language you're most comfortable with, you don't get the premium stuff, you get the slop. Societally, we've decided that. You do what I'm saying? And I think one of the things that's going to continue to happen is that we're just gonna get used to like, oh, okay, well the Slop is normal food.
Dexter:And and we're and for some experiences, we don't you don't deserve the premium experience. There will be people who get the human made stuff. There will be people who get the guarantee that this book was written by a human. That this that this that this journalism was made by a human. But listen, journalists have to eat.
Dexter:Like everybody up here gotta get paid at some point, right? If you're generating slop journalism about stuff that may or may not have happened, that stuff is free. And so the social contract type stuff like every individual having to make the decision that works for everybody in this room I think because I think we're all in the form of like making those decisions for ourselves. What I wanna add on top of that though, is that those decisions are being actively made for us. You do what I'm saying?
Dexter:We're like, again, societally, we've kind of gotten used to the idea that, oh yeah, certain people if they don't, if English isn't their native language, they just don't get the good stuff. That's an unhinged conclusion to come to.
Gita:There's a
Dexter:reason And we decided that's okay.
Gita:There's a reason why, they're pushing AI in really specific categories of media. Pornography, SMUT is like outside of mainstream culture is denigrated usually. Romance novels, even though they are basically single handedly like holding the entire publishing industry on its shoulders, people degrade romance novels and say that they have no no value. I even think the kinds of movies that you saw the Grammys Chinese or TCL whatever, they're I watched some of the clips from them and they're all like big action blockbusters. So movies
Jason:we It's think a budget TV company. Like it's for the people who can't afford like a Samsung TV or whatever. And yeah.
Gita:Yeah. I mean, it's also like the the AI fruit stuff. It's all literally just recreating reality television. Like, it's literally some of the same plot lines from Love Island, like, most frequently. And these are all things that we've kind of all accepted that reality TV is hugely exploitative and terrible, but we also I watch it.
Gita:I watch that shit, like I love that shit so much. You know? I I know that some of these categories of things, like they are all made by humans but we've kind of accepted that some of them are lesser forms and therefore don't need to be respected. But it's in these lesser forms, it's in video games also, you know, it's in animation, cartoons.
Dexter:Who cares? It's cartoons.
Gita:Exactly. Who cares? Kids don't need kids entertainment is also there's a huge amount of AI bullshit. If you let if you let the YouTube Kids recommendation flow just keep going, you will inevitably end up with at least something with AI voice acting in it. And we've just decided, yeah, that those people that like that stuff, those experiences do not matter.
Gita:It does not matter if they're good or not. And
Jason:I think Yeah. I think both of you make such good points and I think we were up here talking about like, you know, specific journalism tools. I was like, oh, we would use AI for this, we use AI for that. That's maybe boring if you're not a journalist. But I think for everyone out there, they need to sort of think like, well, I'm probably going to be asked to use this in my job or in my industry, or I might lose my job to this industry.
Jason:And I think it's hard because it's like a very difficult economy right now, but it's like, to what degree are you comfortable using these tools? And also like as a person, like what do you do better than a robot? Everyone does something much better than a robot. Like I think that the quality, like I guess there'll, as you said, be like a sort of a high and a low in terms of like, well, there's going to be like organic content that is made fully by humans and then it will live in a sea of AI generated content. And I think that standing out is right now still possible, but because like you said, we're not getting to make that choice, the algorithms are making the choice for us and the most powerful companies in the world are pushing this technology on their users, like in ways that often aren't even noticeable or are super annoying, but they just keep doing it and hope that you'll eventually want Gmail to write an email for you.
Jason:And it'll be two robots talking. It'll be two humans talking to each other as intermediated by like just Gmail responses.
Gita:I resent this idea that like a lot of jobs in the future are just going to become an AI minder. Like, that is the the future that is being sold to us, that we're gonna sit next to a machine and just make sure it doesn't go off track, essentially. It doesn't explode on us. Like, we're all gonna end up being Homer Simpson with the drinky bird, you know? Like, that's just it.
Gita:And I think there's more to the life than that.
Dexter:Well, we're gonna have I mean
Jason:Yeah. We're at well
Dexter:Whatever whatever is gonna happen, it's my thing what it comes down to, unfortunately, I'm really sorry it comes down to capitalism. If we want any changes being made, I I have to say that it's probably going to be legislation. It's not gonna be individual people making choices. And I say that because of this. 48% of practicing Christians say that they would trust AI for their spiritual development.
Dexter:And this was this wasn't done by some, you know, non part like, know, whatever critical of Christians. This was made by Christian, specifically a Christian polling, organization. So they would know. What does that mean for pastors, all the pastors in the house, all the priests in the house? I don't know what you'd like.
Dexter:Maybe your job is also at risk too, but I say that to say that if that is if that one of the most inner most parts of humanness within America specifically, right? Myself not being a Christian. There's something a little bit deeper that isn't gonna be solved by everybody, you know, all of us in this room, you know, signing a piece of paper saying no to AI like it just ain't gonna happen. It's gonna be a lot bigger than that, unfortunately, which means it's a lot tougher. But, know, it's good to know what the opponent's move is at the very least, right?
Jason:Thanks so much for listening and watching. This has been the four zero four Media Podcast. We'll be back with another episode in a couple days. If you enjoyed this podcast, please tell your friends about us. This podcast was edited and produced by Alyssa Midcalf.
Jason:It was filmed by the LA Public Library. And, thanks again to the LA Public Library for having us. I'm Jason Kevlar. We'll be back soon.