The 404 Media Podcast (Premium Feed)

They pulled her over in this black Dodge Durango that allegedly hit another car and caused three people to die. And it had not a scratch on it, but they still arrested her. This woman spent thirteen days in jail because she was falsely arrested.

Hello, and welcome to the four zero four Media Podcast where we bring you unparalleled access to hidden worlds both online and IRL. Four zero four Media is a journalist founded and needs your support. To subscribe, go to 404media.co. As well as bonus content every single week, subscribers also get access to additional episodes where we respond to the best comments. Gain access to that content at 404media.co.

Also, remember to subscribe to our YouTube channel where you can watch all of our episodes, including this one. Subscribe at youtube.com/@404mediaco. I'm your host, Joseph, and with me are the other four zero four Media co founders. The first being Sam Cole. Hey.

Emmanuel Mayberg. Hello. And Jason Kebler. Hi. Alright.

So we are having, well, not just one event, two events to celebrate our third anniversary coming up. Sam, what are these two events? And of course, there will be a link directly to get tickets and the descriptions in the show notes. But what are we talking about, Sam? What are doing?

Yeah. We are throwing down in NYC in September. So we're turning three, end of August, early September. So we are gonna throw two events. One of them is a live taping of the podcast.

It's gonna be talks. We're gonna have Becky and Matthew Gault from the abstract and four four come and speak. So you get to meet them and hear them talk about their coverage areas, which are like science and data centers and all that kind of good stuff. So that'll be on September 3 at the Green Space, which is a WMYC event space. It's in, like, the Chelsea area if you're interested.

That'll be September 3. And then on September 4, next night, which is a Friday, we're gonna go to Three's Brewing and have a big party. So come get a drink on us. Come hang out. That's just gonna be like a more casual hangout to celebrate three years.

So subscribers get half off their tickets to Fork Media Live. If you're a subscriber at the supporter level, you can check your membership at forkmedia.co/membership. 50% off the live taping for the party. You get into the party for free if you're a supporter of our subscriber, but there's various ways you can get into this party. You can pay full price if you want to also.

But all of that information is four for media.co/3-years. And you also just go to the website and find it. But, yeah, would be in the show notes, all the info that you need. But we we love doing these. There it's a good chance to, like, meet you guys in person and to see what you're what's going on with you and what you're interested in and and get some good face time.

Even though we're constantly in cyberspace, we do miss seeing you in real life. And we'll all be there. The whole gang will be there, which is a very rare treat. We're never all in the same city. So

Yeah. We're usually only in the same place ever for these parties, basically. Let me just stress, you said come have a drink on us. You can probably have more than one. It's an open bar for several hours.

Yeah. Please have more than one drink. Yes. At Adderberary, drink the beer, please.

Yeah. It's an open bar party for paid subscribers of four zero four Media who get in for free. I don't know like how much more attractive you can make it. And then if you're not a subscriber, you can also just pay for the ticket and you get access to an open bar. But yes, link to more details below.

You should probably get tickets as soon as you can because, I mean, I imagine especially as we get close to it, they're probably gonna fill up. But yes, thank you, Sam. And we'll we'll definitely keep a look keep people updated on that. For this first story, Jason, do you wanna take us through it?

Yeah. So we're continuing to talk about Flock, which over the last few months has really had quite a lot of opposition to it. We we're starting to see a lot of people, like, cut down Flock cameras. I guess there's there's like a lot of discussion about what Flock cameras can and can't do, which is what this story is about. So this story is called How Cops Use Flock to Track People, Not Cars.

Joseph, most people think of Flock as license plate reader cameras, but they do more than that. So what else do they do?

Yeah. So I mean, they absolutely do track people as well and it's not so much specific people like there isn't facial recognition and Flock will always stress that. There's no facial recognition, there are cameras and that sort of thing. You can't type in a name and bring up somebody, but Flock cameras that police have access to can be searched as like a search engine. I think that's how you pull it in your edits to my article, Jason, which I think is a really good way to pull it.

So they can just type in something like male with tattoos, Caucasian woman with UGG boots or something like that or any sort of permutation or combination of these words and they can instantaneously search a bunch of these cameras, sometimes which stretches, you know, into the hundreds and hundreds of cameras at once to look for a specific person. Obviously, even if they're not typing in a name, they're clearly still looking for a specific person because they have this description, right? And I just think that's something that's been really, really overlooked in the flock debate and the flock conversation, and I feel like there's enough evidence to write about it now and like enough dialogue to bring it up as well. But yeah, they are constantly tracking not just license plates that go by, but all of the people as well. And it does depend on the specific sort of camera and we'll get into that.

Yeah. I mean, these are called free form searches. What is a free form search?

Yeah. So a a normal search for Flock is like when you just enter a license plate. Right? There's a little dialogue box where the cop can enter a license plate and that's it. Freeform is a product that Flock released, I think in early twenty twenty five.

And I really paid attention to when it was launched because it sounded very, very interesting. And that Flock was saying, you can now type in natural language searches. Obviously, this is kind of getting on like the AI hype as well, right? Where, oh, I can ask the LLM to do this, I can ask ChatGPT to do this, well, I can ask Flock's system to surface footage that may contain people that I'm interested in. So at the time when Flock announced free form searches and like free form language searches, it was focused on vehicles.

So you could type in stuff like red Ford pickup with a green sticker on it or something like that. Something that is descriptive of the vehicle but wouldn't fit into a license plate sort of search box. And this was how they were advertising it. But as this piece shows and as we'll go into in some of the examples, like it's not just looking for vehicles, you can use those free form searches to look for basically anything. And I just happen to focus on people because obviously, that's probably the most important part when you're talking about a sort of privacy invading technology.

How did you go about finding these? Like, how did you obtain these searches? Was this different than the process of figuring out how Flock is being used otherwise?

So when freeform searches were first announced, again, in 2025, I started doing a bunch of public records requests to various law enforcement agencies who I suspected might soon be using Freeform or might already. I can't remember exactly how, but often when a surveillance company will come out of a new product, they will include a testimonial from a specific law enforcement agency, maybe they were testing it out. Like I've done this with Axon before when they introduced draft one, which uses AI to provide summaries of body cam footage. So I did a bunch of foyers against law enforcement agencies that said they were demoing the technology. I did the same here with Flock and Freeform.

I asked specifically, I want your records and databases related to Freeform searches. For whatever reason, those were denied at the time. So I kind of pulled it to the side and I thought, well, I'll come to this later then when maybe it's more widespread and maybe there are softer targets to send public records requests to and then I'll get like some of these free form searches. Then I was looking around haveibeenflocked.com, which is a very, very interesting website, which basically aggregates the results of a bunch of other people's public records requests against Flock for the license plate searches that we've done to show local cops who are looking up cameras for ICE and tracked a woman who was seeking, who sought an abortion and, you know, all of the data that's contributed to this nationwide conversation on Flock Have I Flock aggregates and maybe not centralizes but puts as much as it can in one place. I was going through there and very handily, Have I Been Flocked already has a section for free form searches because they're actually just included in the ordinary network audit as well.

When we've sent these requests and when researchers who have acted as sources for us, they've done these requests as well, You ask the police force, I want the network audit of your flock searches and that will bring up, here's all the license plates that were searched, this is when they were done, this is how many cameras were queried either in the town or potentially across the country, and this was the reason why. It turns out actually very conveniently for us, the free form searches are just in there as well. I think it's the column text underscore prompt. So thanks to Have I Been Flocked, I was able to just start going through those records and go, oh, okay. There's already a bunch of free form searches out here and now we can start seeing what sort of people searches police are doing flew through flock cameras and not just the license plates.

So what were some of the searches that you found in there? As you mentioned, often when Flock is doing a search or rather when a police is is searching Flock, they'll just type in a license plate. But with this, they can get creative. It's essentially like using Flock as a Google search more or less. So, like, what what are cops typing in?

Yeah. There's kind of like two brackets of searches. Either they are quite specific where it will be somebody searching for, for example, Atlanta Falcons referencing the NFL team. So presumably looking for like a jersey or a shirt with that. Maybe it'll be a cop searching for man wearing a black t shirt and shorts, which I guess isn't super specific, but it still has some characteristics there.

American flag shirt, Dodger shirt came up as well. There's sort of more specific search terms, right? And then you have ones which are just incredibly vague, like tall man, please search all of the cameras for tall man. And look, I mean, I'm sure authorities would say that, well, we have this piece of information that this a missing person or the subject of investigation that we're trying to track down quickly was a tall man. That's all the information we have.

Maybe we'll do that and then maybe we'll get more solid information after that. There's a lot of tall men presumably walking past cameras. So as some of the experts I spoke to, including Beryl Lipton at the Electronic Frontier Foundation, these are probably going to sweep up innocent people as well, arguably even more so than the license plate stuff. Right? I mean, Jason, you you've been tracking this closely.

Like, there's been a lot of mistakes recently.

Right?

Yeah. Yeah. So I wanted to bring up one. This is actually it's a free form search, but it's a free form search for car. So like, as you mentioned, this can be used to track people and we'll get back to that in a second.

But, it can also be used to track cars when the police are not sure of a license plate. So we've seen searches that are just like black SUV in Florida, like things like this, like very vague terms that are going to necessarily catch a bunch of different cars, rather than just like a specific license plate. You can type in like red Acura, and it will will pull up all of the examples that they have of something like that. And so in Florida, this happened in October 2025. There was a hit and run collision in which a black Dodge Durango hit another car and three people died and the driver of that car drove off.

And so the Florida Highway Patrol put in black Dodge Durango because some witness said, oh, it was like a black Dodge Durango. And they pulled over this woman named Lindsey Isaacs, who was driving a black Dodge Durango. And they arrested her and they put her in jail for thirteen days. But she wasn't it was the wrong car. It was the wrong person.

Like, very interestingly and and kind of very insanely, they pulled her over in this black Dodge Durango that allegedly, you know, hit another car and caused three people to die and had not a scratch on it, but they still arrested her. And it turned out that a different black Dodge Durango had hit the the family and and killed them. And so this is just like, this woman spent thirteen days in jail because she was falsely arrested. And so, I mean, this is like an example of a free form search gone wrong. Like they they can do that for cars and then as Joseph has been talking about, they can also put in people descriptors in there.

And so we've seen both types of searches where they're searching like a vague term for a car, or they're searching for, as Joseph said, tool man, like protesters. Like, I've seen searches for signs, like people holding white signs is like a search that we have seen. And so these are like the types of things that can can go in there.

I mean, just because you mentioned like political stuff as well. I contacted the Center for Democracy and Technology, you know, this good organization who have their finger on these issues, they're very reliable when you're like trying to work on a story quickly. Like, please, I need an expert comment on this. And I pinged them about it, and they were already looking to something similar because they sent me a link on Have I Been Flocked to some of the results. And one was very explicit in nature where it said, one police department searched 198 networks for quote, white Jeep with Trump flag, end quote.

Now, I don't know the context like beyond that, but it shows that clearly they can look for a vehicle with a certain political symbol on it, a political flag on it, like no matter what it might be, right? So that was very, very interesting. Again, that was a vehicle search like the one you're talking about, Jason. But then the people searches do show they can like, oh, can look for somebody with a certain branded t shirt, maybe a sports jersey, some something like that. So I I know these searches can get pretty narrow when you're not typing in like black SUV and arresting the wrong person.

Yeah. And and just on that, like like I said, of course, these cameras are networked together as we've reported many, many times as people know now. But sometimes, like, the search is both vague and they're searching a lot of different networks, meaning a lot of different cities or like across an entire state, things like this. And so if you're doing a vague search and you're searching, you know, tens of thousands of cameras, it's going to pull up people who are who have nothing to do with the incident that's being investigated. Joseph, does this mean that like every Flock camera is also tracking people or what did you see in these free form searches?

Yeah. That that that's probably like the one bit of nuance that people really should take away from it because, yes, people should understand the flock cameras are also searching for people and cops are using that based on the search that we've seen, but it's not every single flock camera, right? So ordinarily, when you think of a flock camera and you see it on like a lamppost or whatever or the intersection or wherever you're seeing it, that's probably a license plate reader just in that I think there are more of those cameras, right? These ones are, well, one type they can be are these Condor cameras, which they are explicitly for people tracking, where somebody will walk past them and the camera will automatically pan and zoom on that person because that way that's what it's looking for. These are not strictly the same as the license plate ones.

I think that you can also see this in some of the data where when you look up a search on have I been flocked or if you have the original spreadsheets yourself, it will also say how many cameras were searched and how many networks were searched. When we look at license plate data, when we get that and we've I think we've had untold number of those results at this point, I can't keep track. It is like often, like, 88,000 cameras or something like that or thousands of networks. And when you see that, I feel like you and I, Jason, and you especially Jason, have got a feel for just like, oh, that's a nationwide search because like it's around 80 k, 80,000, something like that. So that shows they searched basically every camera on the National Flock Network for a certain license plate.

That doesn't seem to happen with these free form searches. You'll look at them and like that white Jeep with a Trump flag one. That was a 198 networks, still a lot. That's probably not the entire country, right? Or even some of them are just like a 100 cameras, which are probably Condor or other people tracking cameras installed in that single town or as you say, Jason, maybe in that single state, something like that.

So in the data, it doesn't say the type of camera. You kind of just have to extrapolate that from looking at this data so much over the past, like more than a year at this point. So I would just stress that, yes, Flock cameras are tracking people as we show, but it's not necessarily the same one that's on your street corner. It's kind of hard to tell because they look a little bit different, to the the Condor ones. But we know they track people because throwing back to an earlier story, Flock left a bunch exposed.

Right, Jason?

Yeah. They left a lot exposed. Like, basically, we were able to find the I don't know. It's like a testing situation where, like, a a testing server that Flock was leaving these Condor cameras open on and we did a story back in December on this, that you should check out. We also did a video with Ben Jordan on this, who's a YouTuber who's done a lot of work on Flock.

And basically, like, you know, you could see how the cameras work, where the cameras would follow, people like as they were walking by. I drove up a few hours from where I live and in front of some of these exposed cameras and track myself, I was able to see myself as I was walking around like things like this. And so, you know, Flock has been really marketing the Condor cameras as extremely powerful, like very high resolution. They're called PTZ cameras, pan tilt zoom, So they can, like, basically zoom in on a situation and on people in any given scene, which is like quite concerning. I think the other thing, and I only learned this the other day.

I just missed it when it happened. I don't know if you saw this or not, Joseph. But condors are like the main video cameras that Flock has been promoting and for a while, they were saying that the license plate readers were were just basically like taking static images of of cars that they pass by. But last year, Flock added video capabilities to all of its ALPR cameras as well. And so these cameras are and can film video as well.

They're not as like high resolution as the Condor cameras, and they're not, they're basically like not as powerful as the Condor cameras. They're not really intended to look for humans in a scene, but they are filming video. And, Flock went and sort of told all of its customers that they're pushing this as like a free upgrade to all of the all of the license plate reader cameras out there, last year. And so I would just check that out, if you if that's something you didn't know. I'm not sure if the if they can search people on these, like based on sort of the information that you found.

Like it seems like the number of cameras that were involved in any given search suggested that they're probably Condor cameras because there's not as many Condor cameras out there and a lot of these searches were not as wide as as we've seen with the ALPR cameras where there's sometimes like tens of thousands of cameras in in any given search. But, I think that's something that, people should know that these can film video. And in the, like, promo materials, they basically say, like, you get the license plate, from our reader, but then you can go back and and check the video to see, like, which way a car turned, like, at an intersection, for example.

Yeah. I didn't know that. Yeah. I didn't know that. And, yeah, it's it's it's just very funny.

And and this isn't even a flock specific thing, although it very much is, here, but just like, oh, yeah. We're we're not doing this thing. Surprise. We're doing the thing now, actually. And pushing it as a free update is very, very kind to its customers, but obviously that does set the groundwork potentially for making those into more people search as well if they're high enough resolution, I guess.

Alright. We'll leave that there. When we come back after the break, we're going to talk about Emmanuel's new story about how AI companies are buying a ton of books, it looks like from a particular company. We'll be right back after this. Alright.

And we are back, Emmanuel. I hope I included the correct headline. If not, this this is the one we're going with in the podcast, and I'll just update in the show notes. AI companies are buying tons of old books because they're free of AI slop. So a bookseller you're speaking to, Emmanuel, noticed something happening in April of this year.

What did they see exactly?

Yeah. I guess before I get into that, have you realized that this is a business that operates on this level? Like, obviously, you know, you buy books from Amazon, you buy books from bookstores, but I didn't realize that there are people who make a living trading books at scale. So they'll go to a library that is shutting down or a school that is shutting down, maybe a bookstore that's going out of business and they just buy up the entire stock, then put it up for sale in one of these marketplaces and sit on it for maybe years until somebody shows up and wants to buy it. But that is this guy's business and he's been doing it for a few decades.

And it's not the worst business. It's not a high volume business either. He told me that on a good week, maybe he'll sell 20 books, and that has been the case for years and years. And then suddenly starting in April, he started to get orders for hundreds of books a day. So he will check on his store and various marketplaces and see that he got an order for 300 books, 700 books, 200 books, and that is very unusual.

It is unusual because that is a historic high, and then also the trend has been going in the other direction, he said, because libraries and schools have been getting less funding, not more funding. And when he went online to see if other booksellers were experiencing the same thing, they have these message boards where they talk about their business, and yeah, everybody has said the same thing. They've seen historic spikes in sales.

Gotcha. So what does he suspect around these sales? Like, obviously, as the headline suggests, we're gonna get to AI companies. But like, why does he suspect this is like an AI company potentially or maybe multiple AI companies behind this spike? What's the suspicion there?

So first, it is the volume, right? As I said, it just nobody historically has bought this many books. Then also in the past, when people buy a bunch of books at the same time, they usually follow a theme. So imagine you're a library and you're trying to build out a section on whatever, the Spanish Inquisition, right? Then you'll buy a bunch of books about that subject.

And here, there's no rhyme or reason to the books that are being bought. The only thing they have in common is that they have ISBNs, which is the serial number and barcode on the back of probably every book that you've ever read. And those are the only books that are being bought in these huge batches. And this seller, well as other sellers, he specializes in a certain foreign language and also particular subject, which includes a lot of rare books and a lot of these don't have ISBNs, and none of those books are being bought. So imagine if you're an AI company and you're trying to methodically scan maybe every book in the world, then the easiest way to do that is to just go through a database of ISBNs and go through the list and scan every book.

How does he feel about that? How does he feel about the idea that he might be selling a bunch of books not to a library that's going to serve, you know, maybe kids or adults who are interested in something or even just, you know, a school or whatever? How does he feel about potentially selling books to an AI company?

It's complicated. He it's obviously very good for his business. He says that usually the storage space where he keeps his inventory is overflowing and now there are empty shelves. So it's very good for business. He's making money.

Other people online are talking about and said that they're making money and not complaining. But as someone who's in this business because they love books, there is a likely possibility that the books that he is selling are getting destroyed because the cheapest, fastest way to scan books is not to take a picture of a page, turn the page, take another picture. It is to cut the spine of the book and then feed the pages to a machine that can scan them very quickly. So you're creating a digital copy of the book that is maybe being used for AI training data, but the physical book that is maybe rare, is now more rare because the copy that he sold is no longer in circulation. It's cut and pulped and gone forever.

Yeah. I I guess this is probably the best place to mention what happened with Anthropic before and, like, the lawsuit there and what came out then about them scanning books. Like, what happened with Anthropic?

So we know that AI companies are very hungry for books as AI training data, and we know this because there has been a few stories about them using essentially huge collections of pirated books. Facebook was caught doing this, I believe, and Anthropic was caught doing this. And the reason we know about the physical book operation is a bunch of authors sue pseudanthropic because they use pirated books and training data. And through that lawsuit and discovery, we found out that they had something called Project Panama, which was an mission to scan millions of books and to, yeah, destroy them in the process, then this has been verified by Anthropic itself that that they have been doing this.

Yeah. So I mean, this is hinted at in the headline I read out, but like, why would AI companies want to scan a bunch of books on top of scraping basically the entire web as well? Like, why are they hunting out these books and and scanning them like in in this bulk way? Like, why do that?

So there are two reasons. One is it's just another source of data. Obviously, scraping the entire Internet doesn't sound that easy, and I'm sure it's not, but it's basically been done, right? Like the AI companies have taken everything that's online and use it for training data, and now they're hungry for more training data because that can make the models better. And they're out of like digital material, so you have to go where you haven't been, which is physical material.

Textbooks, printed books, fiction books, anything that you can get your hands on. So that's the first reason. And then the second reason, which is made explicit by one of the companies that is now servicing this effort for AI companies to get printed books, is that we have talked before about this recursive problem with training AI models on AI generated data, which is called model collapse. If you take AI output and use that to train another AI model and you keep doing that cycle after cycle, the AI model starts to get worse and more prone to errors and just generally not as good as if it was trained on human created content. And if you're scanning books from 2022 or or earlier, you're guaranteed to be training on human created content because before 2022, the large language models weren't readily available.

People weren't using it to flood Amazon with AI generated books, and it's just sort of a guaranteed high quality, and most importantly, a source that is free of AI generated slot that these very AI companies are basically producing.

Yeah. The the key date being 11/30/2022, the date that ChatGPT was released. So you're getting a book before that. It's probably not somebody trying to grift the Amazon marketplace by mass producing some AI slot book. Yeah.

That that totally makes sense. So this this brings us, I think, the most interesting part of the piece, which is that, like you found a company that's explicitly facilitating the mass sale of books to AI companies. And like, it's kind of crazy how they just say this all out loud. So what is ISBN DB and and like, what are they offering? What do they say on their website?

ISBN DB has been around for many years. It is sort of like IMDB, is something maybe more people aren't familiar with. It's just a database of all the books that have ISBNs. This has been previously very useful for libraries and schools and retailers who are looking to buy certain books or find out what's in their inventory and organize their inventory and things like this. But since the generative AI boom, they've discovered a new and deep pocketed new client type, which is companies and AI labs that are trying to get these books.

So now they won't only provide you with this database, they will also go out and source these books for you. So they'll find them, buy them with with the AI company's money, and collect them all in one spot and then deliver them directly to wherever the AI company does the scanning. There's a few companies that do this. ISVNDB is one of them. There are other companies that have done this previously for schools or libraries or retailers, but now they are directly offering their services to AI companies that are making these huge, huge purchases, and for them, the the money is is no object because as we know, they they have billions and billions of dollars in investment.

And probably most importantly for these AI companies, especially after the the anthropic stuff, ISBN DB says it can offer these companies secrecy. Right? Like, what does it say about, like, the need to keep this on the down low?

Yeah. So they they talk quite a bit about the optics, the optics problem of doing this, which is nobody wants a headline about their company that says OpenAI or whatever has bought a million books, use them for training data and then destroy them, that just doesn't look good. I don't know what image comes to mind for you, but it's like it's a subject of science fiction and Fahrenheit four fifty one or like book burning and and stuff like that. It's just like destroying printed books is like not a good pretty thing Yeah. To a good vibe.

So part of their pitch is, we'll keep you completely anonymous. This is kind of standard of Discovered for the industry. Right? So it's like if you're on one of these marketplaces where you're selling books, if somebody makes a purchase, you don't know who it is. Right?

You're selling it via the marketplace. It goes to them as a middleman, and then they ship it to the client and the client's identity stays hidden. So they make clear that that is there to protect them, then they also say that they make people sign NDAs. I assume that this is normal as well, but basically, it's like, don't worry. Nobody's gonna find out that you're doing this with us.

That is odd though because like, on eBay, for example, as a seller, like, you know who you're selling to. You get a shipping label. You get you there's like a username associated with who you buy who buys your product. And like, maybe that username isn't identifiable, like, can just be a random username, but you get a shipping label and you know exactly where you're sending it and there's, like, not a middleman. So it is interesting that just, like, I wouldn't consider buying and selling books to be like very sensitive, like under normal times, you know?

Yeah. Yeah. I I believe this is me speculating, but I believe that the reason that this became standard is not originally for privacy's sake. I think it was because purchases are made in the dozens, hundreds, now probably for the AI companies in the thousands of books. And part of the service that these platforms offer is you don't send it directly to the client, you send it to them.

They go through the books, make sure that they meet whatever standard you have, then they repackage them and send them to you. Right? So there's like a quality assurance aspect to this because you're buying such large numbers of books, but now it has and then therefore, like, you send it to an address that's not the final address, but now now it has the added benefit of keeping clients anonymous.

Yeah. And I think just lastly, that, like, this is some of the first English language reporting on this, Right? But there's like rumblings of other stuff. I think I saw some Spanish press, and then I think you throw to a piece out of the The Netherlands. Like, just what what are other countries seeing?

Is it like much of the same thing or like they suspect the same thing or what's going on?

It's the same thing and I can't prove this, but I have a hunch that the reason that we've seen rumblings about this in Spanish language media and now the stories from The Netherlands, and that story says that they talked to a few rare book traders in Europe, and I think that's because the AI companies are going for the rare books first. Right? Like, imagine if there is a race to scan all the books, all the printed books in the world, it would be easier to get like a printed, whatever, Harry Potter. Like obviously there's a digital copy of that as well that's already in the training data, but like a popular book would be easier to get than a rare book. So you'll buy the rare foreign language books first, put those into the database, destroy them, and then you might have like a leg up on the competition.

Yeah. If you are the one scanning and destroying it, then the other company can't scan it because you've fucking burned that book, obviously. Yeah. Alright. We will leave that there.

If you're listening to the free version of the podcast, I'll now play us out. But if you are a paying for a full media subscriber, we're gonna talk about the big Suno hack and everything we learned about that. We were the first to reveal that. The hacker contacted us, and then we published all of these details about Suna. You can subscribe and gain access to that content at 404media.co.

We'll be right back after this. So we're back in the subscribers only section. This is one that Jason wrote. The headline is hack reveals Suno AI music generator scraped YouTube, Deezer, and Genius. Deezer is a name I haven't heard in ages.

So, Jason, you get a tip. I think it was over a weekend. What was it?

Basically, this hacker said, hey, I hacked Suno. Good tip. And I have here's a bunch of stuff from it. And primarily, it was focused on code that this hacker had stolen, from Suno, which is an AI music generator. It's the most popular AI music generator.

There's basically, like, two main ones, Suno and UDO. And Suno is the bigger one, more popular one. And, basically, I got, like, a bunch of messages from this person. They said, I took a bunch of code. I also sent a bunch of, porn to their Slack, which was, I don't know.

No more. Did he send did they send you the porn in the Slack?

I I didn't I didn't get that. But it was basically like it was it was I mean, we've reported on a lot of hacks, and it it was kind of like, here is what I did and here's what I stole. And they took, like they basically, like, broke into a GitHub repository of Sunos and then, you know, had a little bit more access on top of that. But it wasn't like one of these hacks where there was like terabytes and terabytes of information stolen. It was like a relatively limited hack to some degree.

As in like, this person wasn't sitting inside Suno for months. They sort of like accessed their code repository. They logged into their Slack for a bit base basically, like, compromised one employee, and then they were able to grab a bunch of stuff from Stripe, which is like the payment processor. So they got a lot of information about Sunos subscribers. And this was back in fall of last year, but sort of now just kind of like disclosing what they took, you know, in July.

Yeah. Limited data breach in scale of size and scope for GitHub. A GitHub repository is going to be very, very interesting, obviously, for an AI company. I mean, I guess just super briefly, for those who don't know, Suno, like, how is it built broadly and, like, how do people use it? It's an AI music generator, but, like, what does that mean exactly?

Yeah. I mean, so you you basically, like, open it up and it's it's like chat GPT and that you just like are given a window and you type things in and you are like generate me a country song, generate me a rap song about this. Like you can, you know, kind of give it a prompt like you do with any other AI tool and it is made, to generate songs. The types of songs that I most commonly see are this specific meme that has happened on Instagram and TikTok where you take like text messages, from a friend or a boyfriend or girlfriend or your mom or whatever, and you dump them into Suno and you say like, write a song using these lyrics and then that's like become a format. And the lyrics are like the text of the text message.

Obviously, there's like a bunch of other like, people can use it for anything. But largely, like I have seen the way that I have seen Suno been used is like for memes and jokes. There are like tons of AI generated music like on Spotify and elsewhere, but I get the sense that a lot of that music is being done with other tools, maybe like Suno is in the process at at some point. But I think, at least like to my ear, Suno creates music that is, like, relatively easy to detect as AI, whereas, like, I mean, this is a spoiler alert, but Emmanuel's been working on a a bunch of, articles about or, like, a a pretty big article about AI music. And I don't know, like, do you get the sense that Suno is being used for this, or do you think it's like a little bit more hands on than that?

The stuff that I've listened to, I think, is very clearly just one prompt into Suno and up to Spotify it goes, but

I see.

Because it's crappy or what?

Yeah. It's not. Yeah. It's bad. It's simple.

I don't know. Like, there's other stuff that I've listened to where maybe it has, like, several layers of vocals and instruments on it, and it's just, like, interesting and doesn't feel as as generic as the stuff that that Soodan would spit out with one prompt. Yeah.

Yeah. Maybe this is like a little bit of a of a path we're going down that is not super relevant, but I guess I'm thinking of of like Glurb, Glorp? Glorp? Which is like this SpongeBob rapper who who I haven't checked in on in a while, but, like, it is AI, but there is also the person making it clearly has, like, a music production background where they are, like, doing extra stuff to the AI. Like, is pretty heavily AI influenced, but then they clearly, like, know how to use a DAW, which is like a, you know, a digital audio workspace, like a music editor.

And writing, like very clearly written written lyrics and stuff. Yeah.

Yeah. And it's like, I guess the hallmark of Suno and it's like, maybe they are maybe in the long term, like Suno hopes to cater more toward people who are like interested in making complex music, but like, Suno is a text, you type text in and then a song comes out the other side and you're like relatively limited in what you can like do to that song within SunO itself. Like, so it's it's pretty much like you wanna make a song in ten seconds, like, here's how you can do it with SunO.

Sam, just briefly, had you heard of Glaube? I'm googling it. Yeah.

Of course. Yeah.

You're you're usually the one I go to you're you're the one I go to to, like, get a gut check of whether I should have heard of something because Jason is a sicko, Emmanuel's a sicko. You're a sicko, you run away, but I'm also like

my tendency. Yeah.

Well, you're a sicko, but I feel like you're slightly more normie. So like if you've heard you, I'm like, okay. I guess I should have heard you.

Yeah. No. I've heard of Glorb. I the AI music stuff is crazy. Have you heard the it's all over my Instagram reels and like all over my social media in general, just AI generated music.

Have you heard the one that's about hot dogs?

Not me.

Yeah. Emmanuel's nodding. Yeah. He knows. It's

remember I heard it by from you to be clear.

Okay. Okay. I

heard about it. I heard it in my Instagram reels for the first time, and it's so funny to me because I'm 13 years old. But I'm not gonna describe it. You since we're in the subscriber section, I wrote about it in my last behind the blog, and it's linked in there. But yeah.

I mean, it's just like it's really funny. Some of it, like, really gets me. And then I'm like, I feel awful because it's AI. I'm like, man, this isn't actually, like, anything worth anything. But I don't know.

It's it's a brain breaking thing that I experienced on the dailies being out in public being like, is this a Spotify playlist that the person playing it doesn't know is totally AI generated, or is it even AI at all? Like, I don't even know anymore if it is or how to tell. It feels very disorienting.

Yeah. Just quickly, I haven't used Suno in, like, over a year at least, just because I I used it when I did some reporting on it, like, about a year and a half ago, which we'll we'll talk about. And it was very basic then. I'm now looking at it again. It wasn't relevant to the stories, which is why I didn't do it.

But it's like, it seems as though they're adding a bunch of features where it's like, oh, they have a DAW in there now. So basically like the DAW is the thing where it's like, there's a bunch of different tracks and it's like, there's a vocal track, there's a drum track, there's like a bass track, there's a guitar track, like all of this. And you can, like, click and drag them around and edit the music and make it, like, complicated. And Sunoh has started to add some of those features. I think I think they, like, I assume and what it sounds like is, like, you can, like, create a song by typing in a text prompt, and you can say, like, generate a guitar solo here, generate, a drum breakdown here, like so on and so forth.

So so there is like, I guess, additional levels of control that have been added since I last checked it out. But that but, like, generally, like, Suno is just an AI music generator.

Yeah. And obviously, I mean, they've they've said this somewhat and you can obviously figure it out. But to make an AI music generator, you need to scrape a bunch of music from around the web and, you know, there this lawsuit right and that they acknowledged doing a bunch of scraping, but it didn't really drill down into what specifically they scraped and that's where this the information from this hack, you know, is in the public interest because it gives us some insight into that. So these screenshots that you've got, Jason, from the hacker, what did they show specifically about what Suno had or was scraping?

Yeah. So some of the code was basically like the the code that was used for scraping and to organize the databases that Suno was using to train its models. And so there's also there's code, but then there's also comments and the comments were like, you know, in quite plain and obvious English that anyone could understand. So, for example, there's one that explains, let's see. So it talks about like where things come from.

It says Genius HQ, which is like, you know, Genius, the lyrics website, YouTube Music, FreeSound, Jamendo, IMSLP, Deezer, and YouTube Music tagged, which is just like, I I assume is either more YouTube Music or it's YouTube Music that has different, like, metadata in it. And then they specifically talk about how much they took from each individual service. So it says YouTube Music, 113,879 hours. Genius HQ, seventeen thousand hours. Freesound, four hundred and ten hours.

IMSLP, nineteen thousand hours, Jamendo three thousand seven hundred hours, Pond five music sixty two thousand hours, Deezer twelve thousand hours, etcetera, etcetera. And if you like total that up, like if you add the hours divide by 24 to get the number of days and then, know, divide by three sixty five for a year, it's like over twenty years worth of music that has been scraped. And if you didn't know if like, obviously, everyone knows YouTube Music. I think most people know Genius. What I didn't know about Genius is you can stream music on it through Apple Music.

And so, like, it's not it sounds like they were taking both lyrics and music from Genius, which may have then gone on to, like, scrape from Apple Music, unclear. But, like, Pond5 is a stock music website that's run by, I believe, Shutterstock. You know, Deezer is a streaming service like IMSLP is like a sheet music library, I believe, or at least there was some sort of sheet music library on there that was like classical music and things like that. So they were grabbing stuff from like all over the place. And, of course, we like knew this was happening as you said.

Suno was the target of several lawsuits. One of, like, one of which was settled and then several of which are ongoing from the RIAA, which is the recording industry, like something of America.

The association.

Yeah. Association of America. And it's like everyone knows RIAA, like, because they were the all the piracy lawsuits, like the Napster lawsuits, things like that. But then also like, I don't know, I feel like as a teen growing up in like the early two thousands, like the RIAA was like always being talked about both like in a piracy context, but then also like in a parental advisory context, I think. I think that they're involved with that.

I don't know. Anyways, I knew what the RIAA was and they're usually like the bad guy. And in this case, they're like fighting for artists. And so like as the, as those lawsuits have wound their way through court, like, has had to respond to them. And Sunoh has admitted in several of its filings, like, hey, we scrape tons of music.

We scrape music from all over the place. But the, like, exactly from where and exactly how it worked was a question. And it's it's not been clear even from those lawsuits, like, where the music was coming from.

Yeah. So you get this. We do some verification steps, and then we contact for comment. What did Suno say? Because this is one of the craziest one of the craziest statements we've gotten a while.

It was so long and so crazy that we barely included, like, even a quarter of it in the article. We included the important part, which was them acknowledging the scraping, right, and, like, confirming the data breach. But, like, what was the statement that Suno gave you?

Yeah. It was really, really, really long.

It was Did you publish the whole thing in Behind the Blog? I don't think we did. Right?

No. I didn't publish the whole thing. I mean, honestly, like, it's just a very long, long, long statement about Suno's history and what Suno aspires to be and kind of like how, they consider the future of AI music and like all of this sort of thing. I'm trying to pull it up now, but because I didn't include the beginning of the statement, which was like, I found to be quite funny, but like, fully, fully irrelevant.

But the vibe the vibe was like, there was the big bang, and then life moved from the oceans to land. We left the primordial soup, humans evolved, and then Suno was born. Like, that's literally the vibe of this statement. It's crazy.

Yes. The statement is, at Suno, we believe great music starts with people. No matter how powerful technology becomes, music only matters because of the human experiences, emotions, perspectives behind it. Since our earliest days, artists have been central to how we build. We work closely with artists, producers, and songwriters to understand real creative workflows and develop tools that generally support them.

We care deeply about getting this right. More than half of our team are musicians themselves. Then there's like eight more paragraphs.

Yeah, I'm going to cut you off there. Mad boring. Sorry, mad Yeah, boring

extremely boring. And basically it's like they, I mean, they basically don't contest anything that the hacker found. You know, they say that basically, like, the important thing is what Suno is used to do, not where the data comes from. And so they basically talk about, like, how they try to create original new music, not replicate other people. They do admit that they were hacked back in November.

They say they didn't, tell customers about it because they decided that it didn't rise to the level of needing to tell customers.

Even though it had user information and some Stripe information as well? Yeah.

They said, I mean, this part is is important. So they say based on the limited nature of the customer information believed to be involved, we determined that individual notifications were not warranted under applicable privacy laws, which I don't know, that's a determination that they made. I probably would have made a different determination based on the fact that a lot of like payment details were in there, but Mhmm. Yeah. So that's kinda like what they said.

Yeah. I mean, I guess just the last thing I'll say before we wrap up is that that statement about putting humans first and all of that, it comes after those comments, which will be around forever now of the the sooner creator, right, being like, we're making this because I think that people don't actually enjoy making music. So like they're saying that putting humans first and they like fundamentally just do not understand why people make music. And then there was the social media personality who then moved over to I think they were the creator of a fan centric platform, and then they moved over to boosting Suno and everyone called them out for like, what are you doing? I know people hate this company and they have a very, very weird position and they have a very, very weird statement as well.

Alright. Let's leave that there if I can find the script, and I'll play us out. As a reminder, four zero four Media is journalist founded and supported by subscribers. If you do wish to subscribe to four zero four Media and directly support our work, please go to 404media.co. You'll get unlimited access to our articles and an ad free version of this podcast.

Also get to listen to the subscribe zone section where we talk about a bonus story each week. This podcast is produced by Alyssa Midcalf. Another way to support us is by leaving a five star rating and review for the podcast. That stuff really does help us out. I have a few more of those to read out, which I'll do soon.

This has been four zero four Media. We'll see you again next week.