E (1:50)

Depravity.

A (1:50)

Depravity.

B (1:52)

Depravity. Depravity.

D (1:53)

Depravity.

A (1:53)

Yeah, yeah. Has no bottom. I don't know.

E (1:58)

I guess.

A (1:58)

I guess maybe we'll start by explaining what prediction markets are and where they came from.

B (2:05)

I would say, yeah, what the. What the markets are. But then also like the obvious controversies that have come up since they became very popular in the last couple of years. I would say.

A (2:19)

Yeah. So there's two major prediction markets. They're Kalshi and Polymarket. And weirdly, Kalshi seems to be bigger, at least in terms of money spent, at least when I look at things. But I think of Polymarket first kind of. But they are these apps and websites that allow you to basically bet on the outcome of anything that happens, like in the world, essentially. They have taken the model that was started by DraftKings and FanDuel, which are gambling websites that allow you to bet on sports and have allowed you to bet on anything at all. And I think it's kind of important to start like very briefly with DraftKings and FanDuel, because sports gambling has always been a thing. But what those places introduced was something called daily fantasy, where you were able to bet. If you had a fantasy football team, you would maybe pay money with your friends and you would wait an entire year to see if you won or not. Whereas with DraftKings and FanDuel, you were suddenly able to bet on the outcomes of any individual game. And then after that, you could bet on the outcome of. You can bet on hundreds of things per game. And so what it did was it took the metabolism of these bets and made really fast, like you could lose a lot of money very quickly. This has led to. Well, first of all, these companies lobbied super hard all throughout the country to legalize their apps. And they're now legal in most of the country, but not all the country, I believe. And sort of piggybacking off of that has been polymarket and Kalshi, which allows you to bet on the price of Bitcoin. It allows you to bet on sports games, it allows you to bet on who will win elections, it allows you to bet on the price of ram. I don't know if you've seen that, but that's like a pretty interesting one. And then notably, it allows you to bet on the outcomes of various geopolitical strife. So for months now, maybe I guess more than a year, like, you've been able to bet on whether the United States would bomb Iran and when that would happen and what that would look like. And if you bet yes, and the US Bombs Iran, then you get paid out. If you bet no, and the US doesn't bomb Iran, then you get paid out, and vice versa. You can get a lot more specific than just like, will the US Bomb Iran? Um, yeah, yeah.

B (5:03)

Just a couple of things. One is obviously the business of prediction markets in the form of poly market and call sheet is, as you say, an offshoot of sports betting becoming com. Completely legal in most of the country and very popular. I would just add to that that prediction markets have existed for hundreds of years in the form of like an academic, scientific method of predicting things. Usually they're done among groups of experts on conflict or markets, and it's just like one way of gathering information. And the idea is that you get. The theory, at least, is that you get better predictions from experts about what's about to happen if there is money on the line. And somebody basically just took that fairly interesting, fairly useful model and said, let's let anyone bet on the same thing and make money off of it. Which is both, as we'll get into almost definitely a bad business, but also devalues the very idea of what prediction markets are good for. I actually wanted to, before we get into the Iran issue specifically and why we're talking about this today, some examples of how letting anyone bet on anything can go sideways. Matthew, luckily is here. A few weeks ago you wrote about a. I think it was a poly market bet that was related to the war between Ukraine and Russia, and that appears to almost have definitely been cheated in a sense. Can you talk a bit about what happened there?

D (6:54)

Yeah. So people have these really granular bets about how the war is going in Ukraine, between Russia and Ukraine, down to the street level, what sections of the country are going to get taken. So for these kinds of bets, you have to have something that is an arbiter. Somebody has to declare. You have to have some sort of third party that says, like, Russia has officially taken this town square. Ukraine has officially defended this town square. So what Polymarka has been using for a long time is a map that's put out by the isw, which is this kind of independent think tank that literally all it does is it tracks the street level of these conflicts. They were watching one in this particular town and a lot of money was laid on whether or not Russia was going to take it. And it appears that someone at ISW who had access to edit the map, laid a bet and altered the map kind of in off hours to make it appear that Russia had taken a town square and collected their money. That after the money was paid out, because this is all crypto, it's not like, you know, the money can't just simply be returned. Like once the bet was paid out and resolved, the map was edited back. ISW saw this. They did make a somebody was let go or is no longer with isw. But that's just kind of like one small controversy out of myriad controversies.

B (8:33)

This prediction and the lesson being basically, you let anyone bet on anything, you have more ways for people to profit and manipulate the results is essentially the gist of what the issue is there.

D (8:45)

Perverse incentives create perversion, right?

B (8:47)

Exactly. So leading up to the US Israeli attack on Iran and during the protests that preceded them, there's been a lot of bets on the prediction markets about Iran. And one of those was about whether or how long the Ayatollah would stay in power. And obviously he is now confirmed to have been killed in an airstrike. He is no longer in power. But Kalshi is not honoring the bets on that as predicted. Jason, can you talk about why that is and what the controversy is there?

E (9:38)

Yeah.

A (9:38)

So the specific bet was written as Ali Khamenei out as Supreme Leader. Question mark. That's pretty much it. Like, yes or no, and then I guess a date and he was killed in an airstrike. And so kind of by definition, he's out as Supreme Leader. But then Kalshi's CEO and founder, who's this guy, Tarek Mansour, basically said, like, we are not paying this bet out because we don't allow people to directly profit from death. Which is super interesting for a few reasons that I think we'll talk about. But I think that I guess Kalshi just doesn't want to get into, like, an assassination betting pool market, more or less. But basically he argued that, like, the bet was supposed to be about whether the, like, Iranian regime was toppled and not just the Ayatollah. And therefore, like, because the regime is still in power, the bet doesn't pay out. And because he died, the bet is essentially null and void. And I guess, to be fair, I don't think we need to be that fair in this situation. But to be fair, he pulled out some bylaws in the terms and conditions of Polymarket. And it's like, in the event of death, this bet won't pay out very notably. It's like, when you click deeper into these bets, it explains how you would decide who wins under what conditions. As Matt was saying, you need some sort of arbiter of the truth, which is another really interesting thing about these markets. It's like we're constantly fighting about the truth and what ground reality is and that sort of thing. And that's the argument that Polymarket and Call she are making is, like, because there's money on the line, like, we will be the arbiters of truth. Like, as determined by whether these bets pay out or not, more or less. But basically what they said was that winning that bet, quote, requires a broad consensus of reporting indicating that core structures of the Islamic Republic have been dissolved, incapacitated, or replaced by a fundamentally different governing system or otherwise lost de facto power over a majority of the population of Iran. This could occur via revolution, civil war, military coup, or voluntary abdication, but only qualifies if the Islamic Republic is no Longer exercises sovereign power. And then it goes on and on and on like this. And so basically, it's like, if there's a coup, you win. If there is, like, a weird revolution, you win. But it doesn't say, like, if he dies, you win. And to be clear, like, $54 million was bet on this by. By people. And so basically what Kalshee did was like, we're just gonna give everyone their money back and their fees, but we're not going to, like, pay the bet winnings from this. And people have been, like, really mad about it.

B (13:05)

I think it's pretty obvious that this is not Kalshi trying to, like, save money. Kalshi's business model allows for all these payments to be paid out, no matter who wins the bet. They are trying to draw a line in the sand about. I mean, he explicitly said, like, we don't want any bet to allow people to profit from someone's death. When you allow people to bet on active conflicts, that is going to be an issue, even if they do try to draw that line. Can you explain what are some of the obvious problems with trying to frame it that way?

E (13:54)

Yeah, I guess I'll first say that

A (13:56)

the idea of assassination markets has been a thing in cyberpunk for a really long time. And then there's been numerous, like, dark web scandals where there's, like, a kill list and, like, bitcoin tied to, like, if this person dies, you make money. And those are. Those are some of the most, like, dystopian controversial things that could possibly exist. And it should be noted that, like, Kalshi and polymarket are, like, super unregulated in this administration. It's just like, what's happening there is completely insane, but there's almost no regulation of what's going on. And I think that

E (14:37)

what he's trying

A (14:38)

to avoid here is a situation where he, like, people are able to bet on the deaths of other people, and then someone goes out and does, like, an assassination over it. Like, it's. It's some of the bleakest stuff you could possibly imagine. But in trying to draw this line, it's like, well, you can bet on, like, 50 other things about the Iran war. You can bet on things like gas prices, which, if you bet on gas prices going up, that kind of necessarily is betting on this conflict getting worse and people are going to die. In that sense, you can bet on things like, is Iran gonna attack countries in the Gulf on a specific day, like, retaliatorially? And of course, that involves, like, Missiles and bombings and all this sort of thing and people dying. And it doesn't, it doesn't involve the deaths of, you know, the Ayatollah or a president or a well known executive or anything like that, but it involves deaths of civilians and like people in the military and things like that. And so they are drawing this line, but it's extremely murky. And then I think you raised the point, which was a really good one, that was like betting on the Ayatollah to stay in power would be betting on like the deaths of protesters in Iran. Because this, this is a regime that has empowered itself by like simply crushing dissent. And in many cases that involves, you know, quasi judicial killings and things like this and, and you know, executing protesters and stuff. So kind of any way you slice it, it's betting on death. And then I guess the last thing I'll just say, kind of going back to prediction markets being around for a long time, it's like, I guess we should just recognize that the stock market is similar to this. Like commodities trading is similar to this. Like you can bet on like that is gambling to an extent. It's like there's so many people, like Palantir became a meme. A meme stock, for example. And like for Palantir stock to price to go up that requires like horrible, like human rights abuses and expansion of the surveillance state and like all this sort of thing. And so, I mean, there's many layers to it. But I think what happens with like Polymark and Kalshi is just like all of that kind of like artifice has been removed. It's just straight up like, what's going to happen in this war, bro? Let's bet on it.

B (17:15)

Yeah, it's like people are going to die. There's no obvious way to make money off of that unless you're like an arms trader. How can the average person get in on this war profiteering? How about we do it via prediction markets? As you've alluded to that this is sort of a more masked off version of a lot of what is happening in the economy. Anyway, you did come up with, we coined a term called the depravity economy. What does that describe? Like, what is the broader sense in the economy that you think this is an example of?

A (17:56)

Yeah, I think they're not all exactly related, but it's hard to divorce what's happening here from what we've seen happen in the economy over the last decade. Decade where first, you know, you have people betting on Bitcoin, then You have people betting on like increasingly bizarre meme coins that many of which are just like blatantly a scam. You have like weird dropshipping and scams and just like hustle bros doing like lord knows what to, to make money. You have people like throwing their stimulus checks into GameStop. You have the AI gold rush. You have like

B (18:39)

this.

A (18:40)

Yeah. Sports gambling, things like this. And I feel like I don't know where the bottom is. Yeah. Like, I have no idea how much worse this can get, but it feels like this is reaching sort of the culmination where like you can make, you can make like a million dollars if you buy this nft of a, of a gorilla and like flip it at the right time. And now it's just like you can make tons of money if you are super up on the news or have some sort of inside information and you bet on just like death and destruction and human suffering and. Yeah, I don't know, it's just like all vice, there's like, no, all vices are game at this point. I guess it's like whatever you want. Like make money on whatever you want. Like human trafficking, like, who knows? Like, Lord knows what's going on here. And it's just, I think the other kind of weird thing about this is that these markets have been legitimized so quickly because they're one of the few entities in the United States that still has so much money to advertise because their margins are just insane, because they're just taking money from people all day, every day. I mean, they're, they're casinos and so they make tons of money. And so if you listen to podcasts, if you. Not ours, if you hear an ad for one of these, please tell us. Like, if you watch the super bowl, like Polymark and Call she are like on the broadcast. If you go to a sporting event, like their names are on the, you know, the, the court and things like this. And it's just been legitimized because there's so much money here. And then increasingly we're learning, like, Kalsi just signed some deal with the Associated Press. Like Polymarket is integrated into substack and that is now specifically like the business model of Polymarket where they are talking about how like you're going to get your news from Polymarket, which is really crazy.

D (20:55)

That's been Shane Copeland, the CEO. That's been his whole project, if you listen to him talk of Polymarket is that he thinks that Polymarket is a better way to deliver news. To people, it is the future of news. That has been like his long line the entire time.

A (21:12)

So here's their argument for why they are allowing betting on the Iran situation. It's quote, the promise of prediction markets is to harness the wisdom of the crowd to create accurate, unbiased forecasts for the most important events to society. That ability is particularly invaluable in gut wrenching times like today. After discussing with those directly affected by the attacks who had dozens of questions, we realized that prediction markets could give them the answers they needed in ways TV news and X could not. And so they're like, yeah, get, get your news from us. Like we, you will know because we'll either pay out or we won't. And then if you look at their Twitter feed, it's just like a bizarre wire service Twitter feed where they're like, france says it's going to get nukes. Like more nukes. 41% of all scheduled flights to the Middle east have been canceled today. And then some of the tweets are like, new polymarket, is the ayatollah going to be out? They're interspersing this with ways that you can get in on the action of the chaos that we've seen.

B (22:19)

I dropped a link to this website in our work Slack, and I haven't written a story about it yet. If anybody who is listening knows more about it, please get in touch with me. But there's a site called, I think it's Rubit and a lot of people are streaming it and it basically shows traffic cameras from across the world and overlays like markers on the road on top of it and allows people to bet on how many cars are going to make it through this intersection in the next 15 minutes. When I think about the depravity economy, what I think about is the point in mostly American capitalism, our economy, where it's like increasingly there is less real production and more financialization of everything. And that is just like all that is left to do. And it's the only way for people, it's the only way they think they can actually get rich, is to find something that is like crypto, find something that is like the stock market, find something to bet on. Because the idea of like actually working hard and getting a real job that is stable and retiring is like absurd at this point. So people would rather like put it all on black. And you follow that to its logical endpoint. And that's people betting on whether world leaders are going to get blown up or not.

A (23:56)

And that's Nothing of value is being created here.

B (24:00)

Right, Exactly. So before we get off Iran, Matthew also wrote a story with the headline Amazon Data center is on fire after Iranian missile strikes on Dubai. Matthew, how do we first learn that something happened to our precious AWS services in the Middle East?

D (24:23)

People complaining about not being able to access certain services on social media from Dubai. And also all of these videos of as the Iranian strikes occurred in the irreverence in Bahrain, there was a lot of really incredible Instagram reels of influencers like pointing at the sky and watching the missiles crash down. And then AWS has a health dashboard where they publish information about what's going on. And shortly after the missile strikes, I think the first we heard of it was on March 1, they start posting updates about why people can't access certain services in the Middle east region. And a couple days ago when we, or yesterday when I first wrote this story was because they just said that there had been damage to the centers. Unspecified damage, unspecified damage. And I reached out to them and they were like, yes, something has happened. And then they just referred me back to the dashboard. Yesterday at 7 Eastern time, they finally kind of came out with it and were like, okay, it was Iranian drone strikes on, on two data centers in Dubai and another drone strike that landed near one in Bahrain and exploded but wasn't specifically at that data center knocked these data centers off. And it's a thing where like, who knows when those are going to come back up.

B (25:57)

Yeah. So we don't know if either of these data centers was targeted or not. Personally, this is completely me speculating, but it's hard to imagine at this point that they would be targeted. I feel like Iran would have far more high priority targets in the region. But it is a very interesting development. I don't know if people know, but Matthew also has a podcast about conflict called Angry Planet. So the one question I wanted to ask you about this before we move on to the next segment is

E (26:38)

do

B (26:39)

you think countries are looking at a missile strike on a data center? And will that actually, does that change offensive defensive postures for countries around the world? Are we worried now about data centers being charged? Because you can, you can imagine how that would destabilize countries or increasingly as militaries rely on, on AI and big data to like do things?

D (27:08)

What was the big defense story in the week leading up to the attack on Iran? Anthropic fighting with the DOD and are they going to be declared a supply chain risk? Blah, blah, blah, blah. Well, all of like Claude and all of these other services live and die by these data centers that are being built across the entire planet. You know, it's hard to. It would rely on like human or signals intelligence being able to tell you, like, which part of a service is being serviced by which data center. But like, that's a, that's a surmountable problem. You could figure that out. And then, yeah, those are targets. If I can knock out the system that's going to help you do targeting or help you. That's like running missile defense. Like, I'm gonna do it, it'll be the first thing I attack, you know. So, yeah, absolutely. I think data centers are. And it like, causes this kind of civilian disruption like we're seeing here. Like. Yeah, those are, those are targets, for sure.

B (28:11)

Yeah. I think you look at Ukraine and you see like the energy grid constantly being attacked. And it's not hard to imagine in the future how like in the future, in a future conflict, it's data centers because that would be equally disruptive, potentially.

D (28:24)

Okay, well, and they're going to be. There's going to be energy infrastructure built around all of these data centers too, right?

B (28:29)

Yeah.

D (28:30)

Whether it's where we're going to build nuclear power plants and like all sorts of things.

B (28:36)

Okay, we'll leave that there and then when we come back, we'll talk about a story I wrote about Wikipedia. Foreign.

D (28:58)

This episode is sponsored by better health. March 8th is International Women's Day. And so today we are taking a beat to celebrate women and all that they carry at work, in relationships and families, and in the many roles that they play every day. Last year my mom fell at home and suddenly this woman who had supported the family for her entire life, she ended up needing a lot of her own support. And we rallied to her side, but we couldn't do it all. And part of what helped her heal was therapy, the kind of therapy you can get with BetterHelp. BetterHelp's therapists work according to a strict code of conduct and are fully licensed in the United States. And BetterHelp also has a therapist match commitment. It does the initial matching for you so you can focus on your therapy goals. And if you aren't happy with your initial match, switch to a different therapist at any time your emotional well being matters. Find support and feel lighter in therapy. Sign up and get 10% off@betterhelp.com 404media that's B E T T E R h e l p.com 404 Media if

E (30:09)

you've ever felt overwhelmed by the idea of learning a new language? You're not alone. Studies show that 70 to 90% of people trying to learn a new language give up. In fact, I tried to learn Spanish a few times and I gave up the first few times. Then I traveled, felt like I had to learn it, and made a lot of progress. I've been able to both keep up my skills and learn new ways of learning with Babbel, which is built so that it's really easy to get started. Babbel understands that people learn differently, so you can just like dive into a quick podcast when you don't feel like a lesson or you can speak out loud to get that practice in. You can even explore courses based on specific topics and create your own customized review lists all within the app. What I like about Babbel is that it's built around real conversation. The lessons are short, about 10 minutes so they fit into your day, but they're designed by over 200 language experts to teach you words and phrases you'll actually use. I've been using it to work on my Spanish and I've noticed that I'm way more comfortable speaking out loud instead of just freezing up all the time. It breaks things down step by step so you build confidence gradually, whether that's ordering food, making small talk, or just understanding what someone says back to you. Learning a language with Babbel is really about taking small steps and progress. You can actually feel however you learn best, listening, speaking, reading, writing. Babbel adapts to your style and keeps you motivated with real time feedback and progress tracking. They've sold over 25 million subscriptions worldwide, offer 14 languages and every course comes with a 14 day money back guarantee. Make fast lasting progress with Babbel, the science backed language learning app that actually works. Here's a special limited time deal for our listeners right now. Get up to 60% off your Babel subscription at babel.com 404 get up to 60% off at babel.com 404 Spelled B A B-B-E-L.com 404 Rules and restrictions may apply. This episode is sponsored by LED IF and that is if you find yourself using AI personally or for work. You might also find yourself wondering where all your AI conversations go. How are they stored? When are they actually deleted? As we enter into the QUIDGPT era, it's a good time to check up on your AI privacy while searching for alternatives outside of the Silicon Valley surveillance machine. LED was built on the same encryption used by Proton and Signal. You get the power of flagship open source AI models hosted in Finland, all without the techbro surveillance exposure. LED employs a multi layered approach to privacy from end to end encryption to German owned and controlled data centers, all with zero exposure to the US Cloud Act. That's true data sovereignty. Visit Ellydee AI today. It's free to try and their mission

A (33:03)

plan starts at only $5 a month.

E (33:05)

That's LED AI.

C (33:18)

And we're back. This is a story that'll be out by the time this podcast comes out, but is not yet published. So the tentative head right now is AI Translations are adding hallucinations to Wikipedia articles. And this is a story by Emmanuel. So Emmanuel's been following this. The the Wikipedia saga and Wikipedia beef is like a recurring recurring beat for 404 and it's always a delight and never complicated. But you know, the story is no different. So I mean, why don't you just kind of set us up a little bit to begin with with some context on Wikipedia and how it's dealing with AI. If that's something that you can get into like elevator, pitch level explanation, what's their stance at this point? How are they dealing with AI and the influx of generative AI in general, that'll kind of set us up into this AI translations problem.

B (34:20)

I'll try to do it as briefly as possible because as you say, it is very complicated. I'm going to talk exclusively about the position of Wikipedia editors, Wikipedia volunteers, as opposed to the Wikimedia foundation, because it's quite different. But what Wikipedia editors think actually matters more because those are the people who govern the articles you see on Wikipedia. Their position, I think, is extremely nuanced and sophisticated and good and as I've said before, should be a model for the rest of the Internet. But it is essentially that they are not making blanket rules about generative AI because they are open to the possibility of it improving, of it becoming better. They are generally like a very permissive community, but what they are doing is establishing this is my wording, but I think it is fair a kind of zero tolerance for AI generated errors. So they are aware that there's a bunch of generative AI. There is like a special task force of Wikipedia editors that goes around, tries to find generated content with errors on Wikipedia and fix it. And they are implementing special rules on how to edit and fix AI generated content. So that's kind of the general outlook. It's no blanket rules against AI, but high, high vigilance about the errors that it produces because they're already seeing it.

C (36:08)

Yeah, that's fair. And if you're not checking, if you're using AI and not checking it, that's a huge problem and a red flag for you as an editor. So this story focuses on and revolves around something called the Open Knowledge association, or Oka. Can you break down what the Open Knowledge association is and how is that? What role is that playing in this, this AI translation hallucinations problem? Like, what is the problem that we're seeing?

B (36:35)

So I think if you go to Oka's website and read what they do, it kind of looks bigger than it actually is. My understanding after talking to the founder is that it is basically a side project for a Google engineer who's based in Switzerland, I believe. And the idea is that he's a native French speaker. He thought that there were a lot of articles in French that are not available in English, and there's a lot of articles in English that are not available in French. And he's like, I would really love it if there were more translations of articles across Wikipedia generally. Which I think is like a fine and noble goal. But the way that he wants Oka to treat this problem or this challenge is they pay translators, mostly people in the global south who don't get paid a lot of money, a few hundred dollars a month with the expectation of like 40 hours a week for that pay, and explicitly instructing them to use AI to aid them in the translation.

C (38:03)

Okay, gotcha. And you mentioned you talked to the founder. What did he have to say for himself in general? What was his stance on all of this?

B (38:16)

So I talked to him over email over the past couple of days. His name is Jonathan Zimmerman, by the way. And it was pretty interesting because I'm coming to him after wading through very long, detailed Wikipedia discussions. Something that I love about reporting on Wikipedia, there's two things. One, if you're talking to Wikipedia editors, they're the best sources in the world because all they do is write very concise, precise, informative articles. So you'll ask them a question, you'll ask them a question. They're like, here's the information in the right order, summarized as efficiently as possible. So I was talking to an editor who's a source of mine on a few previous stories who first informed me about this. And then the other thing that's great about reporting on Wikipedia is if it's a controversy among editors, the way they govern Wikipedia is that they have all these conversations in the open, in talk pages. And essentially what has happened is that a few people started to notice pretty glaring editors, things like hallucinated citations. This is something that is very common with the LLMs. You'll ask an LLM to generate a scientific paper, it will generate something that reads a lot like a scientific paper with citations, but then the citations don't exist. It's like you'll click on a link to a scientific article and it goes nowhere because it just made up the URL. So there were errors like that, there were factual errors, there were formatting errors because people, these oka editors or translators were copy pasting articles to LLMs and then submitting them as drafts to Wikipedia. The editors noticed a pattern that it was all these okay editors. Then they kind of investigated the company and they found that not only were they instructed to use a AI, not only Zimmerman disputes this, but to them the quota seemed pretty high. Like there was like a high churn of articles that you could imagine would result in more errors. At some point they claim they were specifically instructed to use Grok to translate the articles, which I think is especially a slap in the face to Wikipedia editors because Grok made this Wikipedia competitor called Grokopedia, which is, I mean, I don't know if Jason has a more fair description of it, but it's like it ingested all of Wikipedia and then AI generated an alternative that kind of has like a right wing skew sometimes and then also has a bunch of errors because it's AI generated.

A (41:18)

I would just add that sometimes that is then filtered back into other places and sometimes back into Wikipedia. Not for very long, but there's been a couple instances of that.

B (41:31)

Yeah, it's a real pollutant. And that's actually the reason I wanted to write this, because this is essentially what happened here. You have Wikipedia translators using Grok and then introducing Grok errors back into Wikipedia. Sorry. All that being said, to get to your question, I talked to Zimmerman and his position is the position that I see from a lot of people who are generative AI believers. He thinks that generative AI is useful. It can make people work faster, it can make Wikipedia better. He's not alone in that. I've written before about Jimmy Wales, the founder of Wikipedia also is not as optimistic, but he's at least open to the idea. He says, yes, there are errors, but we have very strict rules about editors having to look and verify that the output is correct. And if there are any errors that got introduced to Wikipedia, he apologizes and says that they need to review and improve their process around that. But he did not say, oh, my bad, maybe we shouldn't use AI in generating Wikipedia articles. Not only that, he also suggested that one new measure they're introducing is having an AI check the output of the AI translations, which, as we talked about very recently with something like AlphaSchool, is a, this point established, flawed method. You can't have an LLM that is prone to one type of errors, check the output of the same LLM for the same errors because it's lacking the ability to catch its own mistakes.

C (43:37)

Yeah, always a great choice to send AI after AI. So they went back and forth, the editors, and they caught a bunch of these errors from Oka. Many of them, like you said, were the source citation errors and just going back and actually looking at the source material and saying, oh, that book doesn't even talk about that page, doesn't even talk about what this Wikipedia article is about. So humans are having to go in and check the work of AI as usual. So where did the editors land on this? What was the final call for them? Maybe not even final, maybe this evolves. But where's the call for them right now on whether or not they're going to let OK keep contributing or OK translators? I guess.

B (44:25)

Yeah. Again, I think the response here from editors speaks to how reasonable, fair and moderate the Wikipedia governance model is. There is no blanket rule against Oka. There's no blanket rule against AI. What they've done is they flagged a couple or not a couple, a few OKA translators who have made mistakes and now they're subject to additional review as opposed to a regular translator or regular editor. And then the other thing is there's like a. I think it's like a two strikes in your out kind of rule. So if an editor uses AI translation twice and they're given a warning, but they find two drafts that have AI generated errors, they prevent the account from making contributions. So I think that's like a very measured, responsible response. Zimmerman is not thrilled about that, but also said that he was part of the discussion and accepts the new restrictions.

C (45:45)

That's good. I guess I also want to point out just AI in translation is a very tricky problem. In general, language is very fluid, has lots of meanings in different contexts, something that AI can't understand in a lot of contexts. It doesn't have the human capacity for language. Obviously it's doing language prediction models and things like that. But a couple months ago, Galt actually published an article about HarperCollins using AI to translate Harlequin novels, which I think is a very interesting use for translation, considering Harlequin romance novels are this very formulaic, prolific genre of book that gets churned out en masse. No offense to Harlequin romance novels and the authors and their work, but it's based on being able to pump through these books very quickly and they're globally popular. So HarperCollins said, you know what? We're going to start using AI in France to eliminate human contractors who are translators who did the work before, which is awful, but is also relying on humans to check the work of the AI. In addition to this. Did. Did this story bring up memories of that story? Because there's a really good comment section happening on our site with the HarperCollins blog and people debating whether. Whether this is a good use of AI or not.

D (47:31)

Oh, they. I haven't seen that debate.

C (47:33)

No, it's not debate. It's like thoughtful, thoughtful comments being made. Not so much a fight.

D (47:39)

It's super interesting. It did bring up those memories for me. It also, you know, and to even further it, like, we saw, I assume, the New York Times profile with the romance authors that are using AI to generate like 40 books a year, which I thought, you know, which I thought was also pretty interesting,

C (48:03)

especially since, like, AI can do, like ChatGPT is going to do erotica. There are so many, like, LLMs that do erotic writing now.

D (48:11)

And those women that the New York Times profiled or also have a side hustle teaching other people how to do it.

C (48:18)

Oh, hell yeah. We're doing pyramid, baby.

A (48:20)

Yeah.

D (48:20)

So it's like their whole scheme is, Look, I'm generating one romance novel in 45 minutes. And on top of that, I am teaching you how to do.

C (48:32)

Selling a class.

D (48:33)

Selling a class. But I do wonder, like,

B (48:39)

what the

D (48:39)

bottom is for that. Like, how many people are actually reading those books, what the bottom is for it, what the market is going to be like in the future, when specifically on art. Maybe people are more interested in having something that's explicitly labeled as this was written entirely by a human being. But I think that, like, romance is especially interesting place to watch because it is so business forward and formulaic that we're. That I think that it's one of those places where any of the advances and the fights are going to happen there first. Right. And then like, all of that will kind of trickle down into the rest of the genres.

A (49:27)

I feel like the market for those books is like, can you trick people into buying them on accident?

D (49:33)

Yep. Because, like, if you've paid. If you've paid a buck 50 or whatever it is, there's a good chance you're not going to fight Amazon to get it back. Or, or Amazon has like the Kindle Direct program where you. Somebody pays like a sub and then they can read as many of these as they want. And if they get, you know, a couple pages into one and see like an AI artifact, like they're just going to move on to the next one, right? And it's maybe already counted as a view for the purposes of whoever's turning out the slob. And I kind of have this wonder in spaces like that if we're not going to get to a place where people abandon spaces that are just kind of filled up with AI and if that may be one of them. And I'm wondering if that'll eventually be a reaction.

B (50:18)

Have either of you ever used AI translation and your regular life, like either talking to an app and having a translate or just like translating some text. Have you ever done that?

D (50:31)

Well, I mean, that's what Google Translate is, right?

B (50:34)

But I bring that up because Google Translate actually I think is AI now, but I've used it a couple of times and it is actually like shockingly good. Much better than the stuff you'd usually get from Google Translate. I used it twice. One is, I think I was translating something for my mom from Hebrew to English and I put it into ChatGPT. I translated it, I read the output, I'm fluent in Hebrew, so I could check it and see that it was good. But it worked very well and she ended up using it. Then I was trying to explain something to my neighbors who only speak Spanish, and I did the same thing. And I asked Jason to kind of look over the translation and see if it was okay. And he actually made some tweaks to it. They weren't like mistakes, but they were like the language was very formal or something like that. And he was like, nobody would actually talk like this. So he made some changes. The point I'm making In both cases, LLMs, not surprisingly, extremely good at translating from one language to another, can be a very useful tool. Crucially in both cases I was able to check it with somebody who actually speaks the language, which is what you should be doing.

A (51:58)

And then it would, theoretically, I use it all the time, like all the time. This is probably the most common use because one, it's like YouTube has an auto translate feature now. So in addition to auto generated captions and transcripts, which are really useful for our reporting, just in Terms of finding things in really long videos. I can now watch an hour long video in Hindi and have it auto translate subtitles to English. And I can tell that it's not that good. Like you can definitely just tell it's not that good. I don't speak any Hindi, but the English that it translates to is very broken often or sometimes it just like can't do it. But for information gathering purposes it's like incredibly useful. And that's how I did almost all of the AI slop reporting just in terms of finding out relevant portions of different videos based on the auto translate. And then we got people who actually speak Hindi to do the direct translations for us. But that's an example. I mean obviously things like translating websites and news articles kind of using the built in tools as well. It's like you, you can tell like if you're translating like a German article to English like using the like translate this entire website, sometimes things are like broken. But it is useful to like get the gist of what's going on and then you know, for your own like information gathering purposes you don't necessarily need it to be 100% accurate. But for something like Wikipedia where you're like directly translating and then publishing like that's, that's not the case. You do need it to be accurate.

D (53:47)

Emmanuel, what did you translate for your mom?

B (53:49)

You said, I believe it was like an artist statement that she previously wrote and needed in English or something like that. Yeah,

D (53:58)

I think that the use case. Right. Is huge because we're talking about for information gathering purposes and being able to have someone check it afterwards. And I think in the instance of romance novels or art and this is a huge thing we haven't really covered, but I've kind of watched that's playing out in anime and the translations of JRPGs. Don't laugh. You're making that face. But it's a big deal. A lot of these companies are moving to auto translated. People are upset because there's something like when you're looking at artistic expression, a translation is not just about a one to one analog. Right. There is nuance in the language and a literal translation may not get across the same meaning in a piece of art. And I think that the LLMs have trouble with that.

A (55:01)

Right.

D (55:01)

The AI.

B (55:03)

Have you ever read two translations of the same book? Two different translations of the same book?

D (55:08)

Yeah. It's fascinating, right?

B (55:10)

No, it is. No, it's good. I love that we're very cultured like that. But it is.

D (55:16)

You're having that fight right now with the. Is it the Iliad or the Odyssey? That there's a new translation.

E (55:22)

Guys, guys, we gotta go.

A (55:23)

We gotta go on tonight. It's like, yeah, that new translation of

B (55:28)

Madame Bovary reads like a modern novel. Jason, you gotta check it out. Okay, let's leave that there. And when we come back to our subscribers only segment, we will talk about Amazon wishlists.

A (55:45)

Thank you for listening to 404 Media Podcast. I'm Jason Kebler. As a reminder, you can subscribe to 404 Media at 404 Media. If you liked this episode, please leave us a five star review. If you did not like this episode, listen to a different one where we have a host who has a real script.

E (56:04)

Please tell your friends about us.

A (56:06)

This episode was produced by Kaleidoscope by Alyssa Midcalf. We will be back in a few days with a new episode. The end. That's the end.

B (56:19)

Perfection.

D (56:20)

Beautiful.

A (56:35)

Lifelock.

E (56:36)

How can I help?

C (56:37)

The IRS said I filed my return, but I haven't.

F (56:40)

One in four taxpaying Americans has paid the price of identity fraud.

A (56:43)

What do I do? My refund, though. I'm freaking out.

E (56:46)

Don't worry, I can fix this.

F (56:47)

Lifelock fixes identity theft, guaranteed. And gets your money back with up to $3 million in coverage.

C (56:52)

I'm so relieved.

E (56:54)

No problem. I'll be with you every step of the way.

F (56:56)

One in four was a fraud. Paying American. Not anymore. Save up to 40% your first year. Visit lifelock.com podcast terms apply.