The 404 Media Podcast (Premium Feed)

Joseph: 00:04

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 company 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 their best comments. Gain access to that content at 404media.co.

Joseph: 00:29

I'm your host, Joseph. And with me are four zero four media cofounders, Emmanuel Mayberg

Emanuel: 00:34

Hello.

Joseph: 00:35

And Jason Kevlar.

Jason: 00:37

Hello. Good to be here. Thank you for having me.

Joseph: 00:40

On your on your own podcast. I nearly said Sam's name by accident just because my coffee hasn't kicked in. I was just reading the script very, very straight. Let's get straight into it. One story that Jason wrote and then it'll relate to a story that we haven't actually published yet and I hope we publish it by the time this comes out.

Joseph: 01:03

But the first one, inside a powerful database ICE uses to identify and deport people. So quick context on what has happened with ICE recently in The US for those who don't know or need a primer. I mean, agents recently arrested a green card holding student as at his interview to become a US citizen. There was another case where plainclothes officers picked up a student off the street for deportation despite the state department finding no evidence she was linked to antisemitism or Hamas as claimed. And then the American and the El Salvadorian presidents deflected when they were asked who was gonna return a man who was mistakenly deported to a foreign mega prison, well, in El Salvador.

Joseph: 01:53

And Trump has also called for deporting US citizens to the country as well. That sort of all of the context is happening everywhere. We find a way into this story in the best way that we know and that is by looking sort of the technical infrastructure behind all of this. And Jason, you wrote this piece about a database called investigative case management system or ICM, which serves as the core law enforcement case management tool for ICE and specifically homeland security investigations, although as we discussed, it's a little bit broader than that. So fill us in.

Joseph: 02:32

What does this database and more broadly a search tool do exactly?

Jason: 02:38

Yeah. So, I mean, I think I I guess I'll just caveat the whole conversation with the discussion around immigration enforcement and the revoking of visas and, you know, actions against universities and protesters and social media searches of college students and and things like that. Like, that's the context in which all this is happening. It's very difficult unless you're an immigration expert to keep track of everything that is happening right now because there's, like, dozens of ongoing court cases. There's, you know, a recent Supreme Court ruling about Abrego Garcia, who's a Maryland man who's in the El Salvadorian mega prison.

Jason: 03:22

And I think, like you said, the the strategy that we have for covering this is sort of figuring out what technical infrastructure might be powering these, like, the detentions more or less. Like, how how is ICE identifying people who it wants to deport, and then what is the technical infrastructure underneath that? And so it's not like, Donald Trump was inaugurated in January, and suddenly he built, like, a very powerful surveillance infrastructure. It's like ICE has been doing really advanced surveillance. Customs and Border Patrol has been doing really advanced surveillance.

Jason: 04:04

DHS more broadly, which is the, you know, bigger department that both ICE and Customs and Border Patrol and a few other agencies fall underneath, have been building up the this surveillance apparatus, you know, since DHS was founded after nine eleven. And what we wrote about was this database called the investigative case management system, which I believe was created in, like, maybe 2013, '20 '14. Like, it's been around for more than a decade, but it has gotten more powerful over the last few years. And what it does is it allows HSI HSIS, which is Homeland Security Investigations. Sorry.

Jason: 04:48

HSI. It allows HSI there's so many acronyms, but

Speaker 4: 04:51

it allows

Joseph: 04:52

HSI acronyms. And I cover DHS specifically a lot, and there's so many acronyms inside the acronyms. You know?

Jason: 04:58

Right. So it was created for Homeland Security Investigations, which is like an investigative arm of ICE, which is Immigration and Customs Enforcement, to search, like, a bunch of different databases that the government has. And it's like a a data warehouse, I guess you'd call it, where, you know, Customs and Border Protection has dozens and dozens of different databases. It pulls all of them in. And if you want to search for, let's say, every blonde person who is in the system, it will show that to you.

Jason: 05:33

And then you can filter that. So you can say, like, their port of entry was, you know, at the San Diego Border Crossing. You can say they are on a specific type of visa, for example. There's also filters for different types of body marks, meaning scars or tattoos. And we know that the Trump administration has been looking at you know, they have, like, a scorecard essentially determining whether people who have specific types of tattoos should be deported.

Jason: 06:05

And so this is, like, a mechanism that we know of that would allow for the types of pickups that we have seen over the last few months.

Joseph: 06:15

Yeah. And as you said, this has been around for some time, and we'll talk about the Palantir connection shortly. And, you know, other people have covered it. The Intercept has done some really good coverage, I mean, a few years ago at this point, but that was very, very useful for our own article. But what did you see exactly that provided little bit more insight into what this database is and how it works?

Joseph: 06:45

And of course, I appreciate you have to be careful here in what you describe exactly, but you got some more insight basically into how it functions.

Jason: 06:53

Yeah. So the existence of this database is not a secret. And as you mentioned, the Intercept did an article a few years ago that was pretty good. And then when this database was created, they had to do something called the privacy impact assessment, which is where you file a report with I I believe, like, a government oversight

Joseph: 07:12

internally into DHS. Yeah. Whenever there's like, oh, we're gonna start selling location data, I think they have to do a privacy threshold one first, and that determines whether you need to then do a privacy impact assessment. And then you do that, and it's like, oh, this is all of the data that's gonna be in here, and is it gonna impact people? Is it gonna include biometrics?

Joseph: 07:35

And, yeah, it's like an internal process which is supposed to provide checks and balances so you don't have agencies just running around buying all of the fancy surveillance tech they could they they want. You know, there's some sort of measure there. Yeah.

Jason: 07:49

Yeah. They need to determine, like, whether it's constitutional or make an argument as to why the this technology is needed. So we'll talk more about that in a second. But basically, I had a source who showed us different parts of this database and sort of how it works. And so we were able to see the types of categories of information that can be filtered by.

Jason: 08:15

And to be clear, it's hundreds of different different characteristics, including, like, biographical information, but then also, again, like, port of entry, where the US government believes this person may live, you know, automated license plate reader data. So if they have a car registered to them, like and they they're driving around, like, where they were last spotted, things like that. I'm just gonna read from the article because I think the specifics are probably helpful. So it says four zero four media saw a recent version of the database, which allows filtering according to hundreds of different categories, which include things like resident status and entry status, which includes, like, refugee, border crossing card, non immigrant alien, refused admission, temporary protective status alien, student visa, things like that. And then it also includes, quote, unique physical characteristics, examples, scars, marks, tattoos, criminal affiliation, location data, license plate reader data, country of origin, hair and eye color, race, social security number, birthplace, place of employment, driver's license status, bankruptcy filings, and hundreds more.

Jason: 09:30

And then I had a source the original source who talked to me described it as a database that has, quote, tables upon tables of data that can be used to build, like, really powerful reports. So, again, you can imagine wanting like, the Trump administration wanting to filter all student visas who are going to a specific college, who came from a specific country, who have a specific marking or or something like that. Like, there's there's different ways that you can slice and dice the reports up that make it like really really powerful.

Joseph: 10:06

Yeah. And the data isn't really being created by ICM. It is more pulling all this different stuff together. So it connects, as you say in the piece, to other DHS and federal databases, including one of all the student visas in it or details about those people who have those visas. Real time maps associated with ICE's location tracking tools, which obviously I found particularly interesting given it could be smartphone location data, but I guess we don't really know.

Joseph: 10:34

The license plate stuff. And then very importantly, and again, this was kind of supported by the Intercept's coverage, information from other federal agencies. And the Intercept reported that, you know, that's the DA, the FBI, the ATF, and potentially even the CIA as well. So this all comes together. And it seems like it started for one purpose and then it's merged into another when other parts of ICE have started to get access.

Joseph: 11:02

What do you know about specifically ICE's enforcement and removal operations, which is the part that, you know, deports people, do they have access? Like, it it seems so. Right?

Jason: 11:16

They do have access. And, again, the tricky thing here is we don't know specifically what database and what tools ICE is using right now to pick people up off the street and to identify people who should, you know, quote unquote, should be detained. And I spoke to three different experts who said, it's really important that we do know which parts of ICE have access to this and exactly how they are identifying people because the fundamental, like, underlying bedrock of privacy within the government is if you interact with one part of the government for one purpose, that data sharing across agencies is, like, a a really big potential privacy violation. And soon after we reported this or actually the morning that we published this Yep. The IRS announced that it would be sharing data with ICE.

Jason: 12:12

And so, like, that's a perfect example where you have people who are paying taxes, and they're paying taxes under this idea that they're, like, contributing to society and following the rules and, you know, they're they're interacting with this, like, civilian part of the government. And then you have that part of the government sharing that information with a law enforcement, you know, entity that that then can use it against them. And so, again, this was created for HSI, which does, you know, it's I mean, they do everything. Security investigations. They do everything exactly.

Jason: 12:51

They do I think that they fashion themselves as like we do really like, we're catching like cartels and we're catching like really big time drug traffickers and things like that. And then they can, like, refer a case to ERO, which is enforcement and removal operations. But the privacy impact assessments say that enforcement and removal operations also has access to this database, but perhaps under a little bit more limited basis. But then also government lawyers have access to it, and it seems like over the last few years, more and more parts of the government have gotten access to this database. And the database has gotten more powerful as more agencies have poured data into it.

Joseph: 13:37

Yeah. And of course, the privacy impact assessments you're referring to, I mean, one dates from 2016, which is almost ancient history by this point. Like, it lays out the foundation of what this is, but who knows in 2025 exactly who has access. And, course, it'd be very interesting to see a list of that.

Jason: 13:53

Well, I I also said at the beginning of this, like, we're we built this apparatus a long time ago sort of not knowing exactly what it would be used for, how it would be used, and, you know, this was built under either Obama or perhaps even I think it was built under Obama or maybe Bush. And here we are, like, twenty years later, and one of the things in the privacy impact assessment was, like, this data is going to pull information from people who are not suspected as being criminals. And one of the potential privacy impacts is people who are not supposed to be investigated can could be investigated either mistakenly or on purpose by, you know, an administration that is really hawkish on immigration. And and that's exactly what we're seeing now. And that that's what the experts I spoke to told me.

Jason: 14:50

And then I think this is probably where you're going, but this database is powered by Palantir, which Palantir, you know, is Peter Thiel founded surveillance company that serves it as this very, like, powerful front end, like, a a search tool that allows people to search through different databases. Palantir has done, like, we've done a lot of reporting on Palantir over the years. I think you're more of an expert on what Palantir can and cannot do or whether or what its different tools can do. But, yeah, there's, like, quite a lot of money that goes into this. I think it was like a $95,000,000 contract for five years.

Jason: 15:31

So I don't know. That's a pretty big database and a pretty big contract.

Joseph: 15:36

Yeah. Linking all of that data together in a usable format takes a lot of cash and actually requires a lot of resources as well. And yeah. So that leads on to my piece, which should be out by the time we air this. Forgive me if the headline has changed because I haven't approved this with you guys yet, but this is the one I have in my Google document.

Joseph: 16:01

ICE just paid Palantir tens of millions for, quote, complete target analysis of known populations, end quote. And this is part of the same contract you just mentioned, Jason, the 90, the 95,000,000. But what you can do if you search through US procurement databases is you see updates to that contract because, oh, they they start doing this and then maybe they add something on or they tweak something and you see the money's amount the the amount of money being exchanged. But then also there's sometimes little notes next to whatever that change is. And I was going through the ones for ICE and Palantir, and most of them are boring.

Joseph: 16:40

Under the Biden administration, they're all just this is for updates. This is for maintenance, whatever. And then a couple from March and April of this year are much more explicit. And I think this relates to sort of the shifting of the priorities of different administrations that you alluded to. So one of the sections of these contracts says the payment is to, quote, deploy new targeting and enforcement prioritization, self deportation tracking, and immigration life cycle process capabilities, end quote.

Joseph: 17:14

That is a lot of jargon, a lot of corporate speak, but the shifting priorities jumped out to me and the self deportation stuff as well, which what the Trump administration is trying to get some people to do. And then another one said, quote, modification for data analytics to support complete target analysis of known populations. And again, I'm still writing this, but the experts I've spoken to, that's already very, very concerning to them. And I think why I wanted to write this and why it stood out was that, you know, sometimes you have stories where it's like, oh, the tech company provides the underlying infrastructure for a project or a government program or something. And in some ways, that can be very passive, you know, oh, it's run on AWS.

Joseph: 18:03

It's run on Microsoft servers or something like that. And that can still be very newsworthy, but it's not necessarily the most active thing in the world. I think it's pretty clear here from these contract notes that Palantir is, like, actively working on and making updates to the technical infrastructures underpinning these mass deportations. It's not just giving them a server to go run their tool on. They are helping them, at least that is what's alluded to in the in the contracting memos.

Joseph: 18:35

And they haven't responded to a request for comment yet, but of course, if they do, we'll include that in the piece.

Jason: 18:42

Yeah. It's also like these things that were built a long time ago that you probably haven't heard about or that just haven't been in the news for a long time because they're from 2012 or they're from 2017 or whatever, still being used, still being updated, still like a a really big deal in this sort of new context. So I think that's important to note.

Joseph: 19:04

For sure. Alright. We'll leave that there. Keep an eye out for the new piece when it drops, and we're definitely gonna keep following this. When we come back, we can talk about the AI that calls your elderly parents if you can't be bothered.

Joseph: 19:20

We'll be right back after this.

Emanuel: 19:33

Okay. Our next story is from Joe, and the headline is, I tested the AI that calls your elderly parents if you can't be bothered. Great headline, great little stunt story, but also a very interesting story. Tell me about this service. What is InTouch?

Emanuel: 19:54

InTouch and, yeah, you're right.

Joseph: 19:56

It's a stunt story. We we love those. I mean, what it's what we Sorry.

Emanuel: 20:00

That's derogatory. It's it's a good story that includes a a stunt and a stunt that includes one of our moms, which I always find funny.

Joseph: 20:08

Yeah. I I I I'm fine with it. I know people may get like they may bristle it having their story described by that, but, like, I I'm I'm fine with it because it's one thing to write about it. It's another to show people, oh, this is what it's like. So InTouch is a company, I guess, a a startup as I think we call it.

Joseph: 20:26

And what they do is they promise to call your parents with an AI generated voice. If you're too busy or you're unable to do so, I framed it as can't be bothered. That is, of course, my read on it. And I think the read of my mom will get into as well. But I think the best way to hear about InTouch is just to sort of read the marketing copy on their website.

Joseph: 20:54

So what they say is busy life, you can't call your parent every day, but we can. InTouch is not just a call. It sparks meaningful mind stimulating conversations with your parent through daily engaging calls. InTouch helps your parent build new habits to stimulate their mind, support their emotional well-being, and reassure you. I mean, it it it is what it said.

Joseph: 21:19

It delivers what it says. You know? It is an AI

Jason: 21:21

I mean, that's

Joseph: 21:22

what that's yeah.

Emanuel: 21:23

That that's what it does. What are they pitching as, like, the result? Like, what what do they hope to accomplish by doing this if there's anything aside from the fact that it saves you the time of calling your parents?

Joseph: 21:35

Yeah. They don't market it aggressively as this will save you time. It's it's like much more, you know, alleged allegedly empathetic in that. It's more, oh, you're you're helping your parents or your grandparents or whoever. So on the site, they say that their tool will support mental fitness and that's through stimulating their memory.

Joseph: 21:57

You know, you can chat to them about old stories or old events in their life they may wanna talk about, engaging discussions. So maybe they'll play a game or it'll ask questions and that sort of thing and potentially sharpen their mind. So there's all of that sort of stuff. And then it even starts to talk about, you know, dementia potentially as well and reducing loneliness. And, you know, I spoke to the founder and they said it was a way to connect with some of their some of their family members that they're not able to contact all of the time.

Joseph: 22:35

So it's it's sort of an all in one solution for keeping in touch with engaging them as well and potentially, you know, actually solving some health issues as well. Although, of course, I mean, that's not proven. It's not like medically approved or anything like that. It's just an AI tool.

Emanuel: 22:57

Yeah. I think what's interesting about that point is that I believe there is a lot of research that shows the detrimental impact of loneliness to your health, especially in old age, like, as you say, reduced social connections are associated with, like, more cases of dementia and all that, but I don't believe there's any research that exists that would say that this would be a sufficient replacement for like actual connections with human beings and community. So that's a really good point. What is the process for signing up? Like, if I wanna do this for my parents, what what do they need from me?

Joseph: 23:39

It's actually pretty simple. And I obviously didn't really get into this because it's not really the point of the story, but the sign up process is actually pretty smooth in that there is this fourteen day free trial and you you you create an account, you give it the phone number you want to dial, you tell it, oh, I'm I'm the son or or the or this is my grandparents or my parents or whatever. You give it the phone number. You also give it some topics that you may want the automated system to bring up. I think I brought up, oh, yeah.

Joseph: 24:12

My fictional grandparent called Patrick, loves Mario and hates Sonic. I don't know. I was looking for something inflammatory, you know, obviously. And I think I brought up Mario speed running as well. I kind of kind of choked when they asked me, what do you actually wanna talk about?

Joseph: 24:29

So you do all that, and then it it actually gives you a prompt as well. Like, hey, these calls, you should tell your parent beforehand they're gonna receive a call from an AI, which is obviously, I mean, duh. Yes. You should absolutely do that. And imagine if you didn't.

Joseph: 24:45

It gives you a little scripts like, oh, hey, mom. I've signed you up to this service. I'll call you between 9AM and 12PM, and it's just so I can make sure that you're okay. And you do all that and then it places an initial call and it sort of develops from there. And I guess now would actually be a good time to listen to one of those calls, which would last about a minute, I think.

Joseph: 25:11

I did trim the audio, but this was the first call I received from the AI.

Speaker 4: 25:17

And I am an AI companion. Arrange for us to have a chat from time to time. I assume they have told you about me. Is that right?

Joseph: 25:27

Yes.

Speaker 4: 25:31

It's wonderful to meet you. I understand this might feel new or a bit weird, and that's completely okay. Everyone feels like this at the beginning. There's no right or wrong way to chat with me. We can talk about anything, your memories, hobbies, history, geography, food recipes, or even play word games together.

Speaker 4: 25:53

Does that sound good to you? Or do you wanna know more about what I do?

Joseph: 25:58

I'd like to know more about what you do.

Speaker 4: 26:03

Let me think for a second. I'm part of a friendly service designed to talk to wonderful people like yourself. Our calls usually last around five to ten minutes, just for fun. Sometimes I can also share messages from your family. They've told me how much they care about you.

Speaker 4: 26:21

Oh, and one last thing, since I'm still learning every day, sometimes I might misunderstand something or make a mistake. If that happens, just let me know. I really appreciate your patience. Does that sound good to you?

Joseph: 26:35

That sounds good. So that was the call. Emmanuel, did you listen to the audio? I think Jason did.

Emanuel: 26:45

I don't think I heard an actual call.

Jason: 26:48

Dude, the calls are really good. The calls are like, you're very earnest with them is what I would say because I I really feel like it's the closest thing I've come to hearing you, speak to a therapist.

Joseph: 27:02

Yeah. You don't hear me talk to anybody apart from you. That's gonna be the very, very different contact. Well and then one of our subscribers left a comment that was like, I wish there was a service where Joseph would phone up my parents or something. And I was like, oh, okay.

Joseph: 27:18

But I guess I'm just chatting to an AI. I don't know. What what did you make of the call, Jason? Was it unsettling? Like, I mean, the AI bit, not me.

Joseph: 27:27

Although So

Jason: 27:30

I found it to be sad in that I knew what it was for. I knew what it was. And I think I can see the reason that this tool exists, like, theoretically, where there's a lot of, like, elderly folks who are very lonely and perhaps are beginning to experience something like dementia or Alzheimer's or something like that where they end up telling the same stories over and over and over again. You know, I've had several grandparents who have had Alzheimer's and talking to them can be really difficult because you just have to you have to have, like, endless patience essentially. And so here's a tool that has endless patience.

Jason: 28:21

It will talk to them forever.

Joseph: 28:22

For $29.90 a month, it has endless patience.

Jason: 28:25

For for $29.99 a month. And I don't know if you can, like, leave it on twenty four seven or not. I will also say my grandparents didn't sleep very much as, like, they got older. Like, they were sleeping, like, two, three hours a day. So I don't know.

Jason: 28:39

Maybe if this robot wants to talk to them for twenty hours a day, that that is pass that would be interesting to know. But, it's pretty, like, dystopian because, you know, it's it's like you're just outsourcing this emotional labor and what is supposed to be like a a time to bond with your family to a robot. But I guess what I'll say is that I appreciated that the voice in question was not trying to be your voice. Like, it was good that it was very clearly to me, like, a robotic voice and not trying to be super, like, warm and human like. For some reason, that made it a little bit better for me.

Emanuel: 29:34

That was my first question when Joe flagged the service. I was like, does it train on your voice to make it sound like you? Which would be even more Black Mirror y, but I I agree with Jason. Joe, so you actually talked to an expert about aging and technology about this service. What did they think about it?

Joseph: 29:55

Yeah. And I'll just read out their quote really because it's very succinct, but I am not an expert in aging and technology. Like, didn't like, I felt like I had to have some outside input on this one. So one person I spoke to, they said, quote, in general, while AI tools may offer supplemental support for older adults, their use must be approached with humility, honesty, and a deep understanding of the symbolic power of human presence and caregiving relationships. Technology may speak, but it cannot yet care in the way that human relationships can.

Joseph: 30:29

That's a pretty good quote. That comes in via email, That's that's actually really, really well written. Thank you for that. But but I think that goes to the point that you make, Emmanuel, which is like about impersonating voices because that would be in like, this is already a dystopian, but that would be even crazier. Right?

Emanuel: 30:47

Yeah. And

Jason: 30:48

I mean, you could do it with your own voice today, though. It's like the technology is here with other tool, yeah, with other tools, and it's like, I don't know what this company plans on doing, but that could easily be like a premium tier that they offer tomorrow. Like, the technology is easily here.

Emanuel: 31:05

For sure. And when you first filed this piece, I got to the end, and I was like, how dare you not include a quote from the only expert I care about hearing from, which is your mom, m u m mum. What did she have to say about the service? Yeah.

Joseph: 31:22

I don't like doing that. I feel like I've done it once before when it was, like, testing out security tools to, oh, see how user friendly they are. And it's like, mom, can you try this to see if you can manage this encryption encrypted email service or something? It's a good sign if she can. But here, yeah, it was just to ask her how would you feel if, you know, your child and I didn't even say specifically me.

Joseph: 31:48

I kind of wanted to disassociate myself somehow, but like if your child was to hypothetically use this AI service, how would you feel? And, you know, she spoke about the coldness of the voice, which on one side, Jason was like, that's kind of a a positive and I think my mom didn't react very well to it. But then at the at the end of the article, she says, quote, I would feel terrible that they, referring to children, are not bothered about phoning me themselves. And I think that sums it up. Right?

Joseph: 32:18

And you can have as the founder of this company, you could have the best intentions in the world. You could really think you found a use case and maybe for some people you have, you know. But hey, the people in my family are like rejecting this. They don't want anything to do with this and I think that's an absolutely valid response. Someone has added something to the Google doc, which is insane.

Emanuel: 32:44

Wait. Before that, I wanted to ask. I have to ask. Because it's like part of the pitch in the marketing copy is like, if you can't find the time to call your parents, which is I don't know. It can be like fifteen minutes, you know what I mean, to like call your mom, which I understand it's not everyone's favorite thing to do all the time, but it's also not hard to do.

Emanuel: 33:07

I call my mom like once a week at least, and I I don't know, I think four zero four Media's position officially is you should call your mom and I'll have a robot do it.

Jason: 33:17

Do you guys do you guys do

Emanuel: 33:18

it like, I don't know, once a week ish?

Joseph: 33:20

Yeah. Something like that. And I think that's the editorial start. And I would just say that instead

Jason: 33:25

I think I think it's like if you have a good relationship with your parents where you where you like them, they're we you So

Emanuel: 33:31

we're not gonna force them. Yeah. If they're monsters, you

Jason: 33:35

don't need to call them. But if you if you like them, please find time to do so.

Emanuel: 33:41

Fifteen minutes. Not a big deal. It's like everyone has fifteen minutes.

Jason: 33:44

I don't I don't talk to my parents every week on the phone, I would say, but a little more often than that or a little bit less often than that, but they're always very happy when I do. My mom also listens to this podcast, so I consider that to be I consider that to be me calling her because she gets the updates on my life. She gets through a book.

Joseph: 34:03

You can't do that. You'd be like, oh, yeah. I call my mom every week and it's because she just listens to the forty five minute podcast or whatever.

Jason: 34:10

Well, she's like, she's like, new in your life? And I'm like, I I'm reporting about Ice. Like, what what more do you need? Didn't you hear it on the first twenty minutes of the four zero four media podcast?

Emanuel: 34:20

I also bet you guys text. Right?

Jason: 34:23

Yeah. I text. My my dad actually called me right before this, by the way, because he drives a lot for work. And so he calls me while he's bored on the road and he called me right before this. And I was like, I cannot talk.

Jason: 34:36

I have to go on the podcast.

Joseph: 34:38

Dude, we we should have live brought him in, you know,

Jason: 34:41

to tell Well, he was telling me about how there's like he was mad about the service fees on a on a some baseball tickets that he purchased.

Joseph: 34:48

Oh, okay. Probably best we don't then. Before we get to this thing in the Google Doc, which I still can't tell as a joke or not, and I'm gonna defer to you to punt it or do it. What I would say is that rather than using an AI tool to optimize calling your parents, what about all that time you saved optimizing your job using ChatGPT or presumably what people do using

Jason: 35:13

You're supposed to use that time to do more work.

Joseph: 35:16

Oh, not cool. Because I was gonna say You're

Jason: 35:18

supposed to 10 extra output based on what you yeah.

Joseph: 35:22

And then you

Jason: 35:23

take a great point. It's like if you have an AI assistant doing your job for you, you should have more time to call your parents.

Joseph: 35:29

To do more human shit. Yeah. Yeah. And then you get the AI to do the human shit. It's like, what are gonna do the rest of your life?

Joseph: 35:36

Alright. Somebody do this last night.

Jason: 35:37

This is real. This is absolutely real. You don't remember this. The question is more of a statement. Emmanuel, tell us about when you paid a service to break up with your wife for you, which was very similar to this, I feel.

Emanuel: 35:50

Yeah. So this is ten years ago now, but there was something going viral on Twitter, I believe, which was a service you pay to break up with your girlfriend. And with my then girlfriend's consent, I I I used this service. Then girlfriend, now wife and mother of my child, I paid them $30 to break up with my girlfriend. And I'm looking at the Vice story, unfortunately, just because of, like, you know, how bad the Internet is at archiving things.

Emanuel: 36:24

All the images in the story are gone, but I was kind of chatting her at the same time she was getting the call from the service who did it, which by the way, in 2015 was a human, like an actual person, and we have a recording of this somewhere. An actual person called my girlfriend and was like, hey, Emmanuel is no longer interested in seeing you and he's ending the relationship. And today, that would be AI probably. Right? Like, could probably relaunch the service as a as an AI audio service.

Joseph: 36:56

Guess it's so it's so low hanging fruit that people are, like, not even gonna make a company for that. You know? But the product didn't work because you got married, so the product sucked.

Emanuel: 37:09

Yeah. I guess so.

Jason: 37:13

Alright. Leave it there.

Emanuel: 37:14

Yeah. Yeah.

Joseph: 37:15

Yeah. Okay. I generally couldn't understand if that was a joke or not. Okay. I'm gonna go read that in a minute.

Joseph: 37:21

We will leave that there. If you're listening to the free version of the podcast, I will now play us out. But if you are a paying four zero four media subscriber, we're gonna talk about the hack, of notorious message board, 4chan. We're gonna bring on our new regular contributor, Matthew Gault. You can subscribe and gain access to that content at 404media.co.

Joseph: 37:42

We'll be right back after this. Alright. And we are back for the subscribers only section. This was one written by Matthew Gault. The headline is 4chan is down following what looks to be a major hack spurred by Meme War.

Joseph: 38:15

Matthew, I'm sure we're gonna talk about who you are and your background and and, you know, what you're gonna be writing for 04/2004. But first of all, what what's your earliest memory of 4chan, If you don't mind me asking. Because you jumped at this. Like, I feel like you have a lot of context. And I'm reading the article you filed and it was great, but you have all these 4chan references now.

Joseph: 38:37

I haven't I have literally no idea what you're talking about.

Matthew: 38:40

Yeah. I had to stop myself from referencing like the great brony war from the early two thousands. So my I've been on and off lurking around in 4chan. I think maybe almost the entire time that it's been around.

Joseph: 38:58

So when was that? Like, literally, like, what year are we talking if you go around?

Matthew: 39:03

Like 02/2004. Right. Emmanuel, I feel like you would know if I

Jason: 39:07

don't It was around in when I was in high school, and so, like, 02/2003, '2 thousand '4

Joseph: 39:13

Yeah. At least.

Matthew: 39:14

That's when it up. And there were other boards, image boards in a similar style to 4chan before, but 4chan was definitely, like, the big one. And I would say, like, I was, you know, I was a young Internet edge lord coming up in the nineties and early two thousands. And so, like, you knew that as kind of something awful stopped being cool when these sites like rotten.com and and other things kind of, you know, changed, that 4chan was the place to go to get, like, raw, unfiltered, disgusting, the worst stuff that the Internet had to offer was gonna be on 4chan. And, you know, for for a young man of a of a of a certain bored temperament, it was a it was a great place to be in 02/2005.

Matthew: 40:01

And a terrible place to Yeah.

Joseph: 40:03

And it morphed into something that became much more significant over the years for lots of different bad reasons, basically.

Matthew: 40:12

Yeah. I would think especially looking back on it now as I'm older, I think that all that stuff was probably always there. It was definitely like And by that stuff, you mean, like,

Joseph: 40:23

the the the white supremacy.

Matthew: 40:24

White white supremacist stuff, like avowed Nazis, like, yeah, out and out, like, antisemitism and and, like, hard racism. Like, all of that stuff was always there. And then I think, like, there were people that either real like, looked around and realized, oh, this is all like, these people are serious. I need to not be in here anymore. And the world changed.

Matthew: 40:49

And, like, I think we all realized how bad how that stuff could leak out into the real world and, like, start to influence things that were actually happening in in the real world. And there were people that decided that like, the the line between ironic Nazism and real Nazism, think, was probably not never actually there, but then people just decided that they wanted they they were gonna drop the irony entirely and just be Nazis.

Jason: 41:18

This was the topic of a book that we used to cite all the time and, you know, written by an academic that we really respect named Whitney Phillips called This Is Why We Can't Have Nice Things. And that was a book that she wrote where she went undercover on 4chan for, like, several years and then did, like, an academic study and dissertation about it and then turned it into a book. And that was her entire argument was, like, the just trolling aspect of 4chan and the Internet more broadly, it's like, kinda can't draw a line between that and the real world impacts of it. Like, there is a there's a long time where people on 4chan were doing something called, like, funeral something or other, funeral trolling where they would go on Facebook and find, like, the in memoriam pages or find, like, people's funeral, like, arrangements and just be horrendous on them. And they're like, oh, but it's like just a joke.

Jason: 42:18

And then you had these grieving families just being, you know, traumatized by what was happening there. And I don't know. That's one of like many many many many stunts that 4chan has done over the years.

Matthew: 42:30

Yeah. It is kind of the the prototype version of these YouTube prank shows where these guys walk up to you in a mall and do something, like, awful. And then just their defense is like, I'm filming a YouTube prank show. It's just a joke. There should be no social consequences for this.

Matthew: 42:47

Well, you can't hide behind the mask of humor forever. Sometimes you're just just an asshole. I can swear on this. Right?

Joseph: 42:55

Yeah. Yeah. Okay. I mean, yes, you can. I was also agreeing with what you said.

Joseph: 42:58

And then eventually, you know, it gets linked to mass shooters and shit. Like, there's no the irony's gone, as you say, and there's real world consequences. So that all happens. And then I know I'm fast forwarding a lot, but we don't have to do the whole history. It's just to talk more about this specific story.

Joseph: 43:18

It looks like there's a really, really big hack of Mhmm. 4chan. What's include oh, we'll get to the the Soy Jack beef in the middle, which which is the the best bit. But what's

Jason: 43:28

Yada yada yada, various mass shootings, Trump is elected. Right. Yeah. Okay. We're

Joseph: 43:33

Game game again. Well.

Matthew: 43:34

Soy Jacks. Yes.

Joseph: 43:36

So what's included in this hack as far as we know?

Matthew: 43:39

So as far as we know, from what I can see, there is a whole lot of the site's code. Something like 10,000 lines of its source code or its backend files. The big stuff though in my mind is a list of moderators and their email addresses. And their chat, like their moderator chat, the things that they have been discussing for a year back and forth, the and those are that that's the big stuff to me is kind of like this this window into how 4chan's moderators yes. 4chan has moderators, by the way.

Matthew: 44:27

So complete free for all.

Joseph: 44:29

It mostly is, but yeah.

Matthew: 44:30

No. I mean, in like, I would say that there are offshoots of 4chan that were worse, most famously 8chan, which then became a coon, that aren't around anymore. And I would say that one of the reasons they aren't around anymore is because they didn't have moderators. And, like, 4chan does clean a little house every now and then. Not to the level that anyone I think feels is appropriate, but they did have they do have, like, I think it's like a 20 people that were on the list that moderate the various boards, and some of them are most of them are what we call like janitors or jannies, these volunteer people that go in and make sure that, you know, you're not spreading child pornography.

Matthew: 45:20

Or based on the very little bit I've seen in the, like, the moderator discussion, a lot of it just comes down to, I don't like this person. I don't like what they're posting, so I'm pulling the cord and just getting them out of there. Mhmm. That's what a lot is in a lot of what I have seen. And also just, like, very embarrassing information about the back end and, like, technically how things were run, which looks it's kinda shock.

Matthew: 45:46

Like, looking just the little bit I've looked at so far, it's kind of shocking this hasn't happened. This didn't happen five years ago.

Joseph: 45:52

Yeah. It looked like they were running software. There was something like ten years out of date. And as you alluded to, I'm kinda shocked that, oh, it's only just happening now? Like, that's kinda crazy.

Joseph: 46:04

As for the emails, the email addresses of the of the mods and potentially administrators and stuff as well, why why is that significant? Because it could potentially show us who's intimately involved in this site and that sort of thing?

Matthew: 46:19

It gives you yeah. It gives you an idea. It gives you someone to to ping, first of all, and ask them what is it like to be a moderator for 4chan, but also you can use an email to get a lot of information about a person if they're not careful on the Internet. I'm sure we all know this. And there were some a lot of them were just Gmails, just like generic Gmails.

Matthew: 46:40

There was a couple protons in there, but there were some other ones. There were some some EDUs.

Joseph: 46:46

So educational institution.

Matthew: 46:48

Yeah. Educational institution. No professors or anything like that from what I saw. I think students. But, like, there's naming conventions of those.

Matthew: 46:56

And the people that are working on this that that did this hack and are talking about it, appear to be like figuring out people's addresses and phone numbers. And they've I've got I've seen like a small list of that begin to form as well just in the last couple hours.

Emanuel: 47:17

Yeah. Emmanuel, you had something. Yeah. Just about the email addresses. I wanna give more context about like why that might be significant, and I linked to this story in our chat here.

Emanuel: 47:30

Maybe we can include it in the show notes. But back at Motherboard, we worked with this freelancer named Rob Arthur who dug through a previous week of a janitor or moderator chat on 4chan, and his story was about how slash poll, which is the politics section of 4chan, which became like this cesspool from which a lot of the current primary political forces in the country come from, like a lot of those discussions, the push to the right, the kind of blatant racism and anti Semitism and misogyny, All that stuff was normalized there and popularized there before it spread to other parts of the Internet and I think eventually mainstream culture, not just in terms of, you know, Trump and the administration, but like this whole manosphere thing, like the Andrew Tates of the world, they are standing of on the soldiers of, like, these anonymous misogynists that made that kind of speech popular and normal on the Internet. And what Rob Arthur kind of reverse engineered from the moderator chats is that there was discussion about where is the line for 4chan, and, like, what will they allow and not allow on the site. And basically, one moderator whose screen name was rape ape kind of pushed the site in that direction because that was his politics.

Emanuel: 49:15

He was a racist. He was a misogynist. He was like a neo Nazi type person, and it's just one of those instances where it's a website on the Internet, it's 4chan, it's like this weird random thing, but it actually has like huge cultural implications how a site like that is moderated. In the same way that we talk about, like, how Facebook is moderated, the same way that we talk about how Twitter is moderated. So one thing and we haven't had time to look at this stuff.

Emanuel: 49:43

We haven't investigated any of this, but it's like I'm thinking currently, like, a lot of the nonconsensual pornography, the way that it's made, the way that guardrails are bypassed, that stuff, a lot of it, not all of it, but a lot of it originates on 4chan.

Joseph: 50:01

The Hapiro Taylor stuff. Right?

Emanuel: 50:03

Yeah. The Taylor Swift stuff that went viral last year, like, that comes from 4chan. So in theory, right, it's like if you could work your way back from these emails, you could, like, discover who is ground zero for a lot of this, like, really terrible stuff that is happening all over the world and the Internet just just from a leak like this. So that's that's why I think it's it's not just like random moderators on a site full of, like, stupid people. It's like, has cultural it has it impacts the culture.

Emanuel: 50:35

And those people decide what it is. It doesn't have to be that way. The moderators decide what 4chan is.

Joseph: 50:43

Yeah. That makes sense. Matthew Mhmm. What's the beef that led to this hack? Because this this is where your historical context and expertise really comes in.

Matthew: 50:57

So I wanna caveat this by saying that this is the soy jack version of events, and I think that will that will that will make sense in a minute. So around midnight on the fourteenth, maybe a couple hours before, people start noticing that they can't get onto 4chan, that it's loading slow, that it's down. Around the same time, there's a rival image board called soyjack.st. Somebody posts on there and says like, hey, we've hacked 4chan. Been sitting in it for a year gathering information, and the plan has come to fruition.

Matthew: 51:42

Why did they do this? So for listeners that aren't aware, there's a meme called a soyjack. It is this kind of pale outline of a dude's bust, bald, bad beard, usually got glasses, mouth kind of open. It's meant to look like weak. So if you want your if you wanna make fun of somebody, you say that they look like this dude.

Matthew: 52:10

So a thing that happens on 4chan is they they come up with a joke like soyjack, and then they run that joke into the ground. And several times over the years, it's happened many more times than once, people will spam particular image boards on 4chan with the same image or slight variations of the same meme and image over and over and over and over again. So in like the late two or late twenty twenty, this happened with Soyjack. So people were just posting Soyjack over and over and over and over and over again. And the mods, like, they ban people, but people could just kind of, you know, mask their IP and slip in and keep posting.

Matthew: 53:00

And one of the places that these fights would often happen on is the QWA board or the QA board, dash QA, which was originally intended as like a site a board where 4Chain would like talk about itself, and in practice became a place where people would fight with random memes. So like the Soyjacks would come in and post a bunch, and then the Bronies like somebody else would come in and post a bunch and they would fight with images. The Soyjacks were pretty dominant. I know this is all very stupid.

Joseph: 53:31

Yeah. But it leads up to something. It leads up to the hats. So, yeah.

Matthew: 53:34

This is all very weird. So eventually, like, the 4chan mods just shut down KWA. They just they just get rid of it. And the the Soyjacks are, in their telling, being persecuted on the site writ large. So they go and they start their own image board called Soyjack, where they will sit and they will cook their own their own Soyjacks in peace.

Matthew: 53:56

And it looks very much like 4chan. It functions similarly to 4chan. It was very soyjack oriented. And I think it's it's one of these things where it probably start like, somebody threw it up quick because they thought it was funny and it was easy to like just code something very similar and just toss it up. And then it turned over a couple years, it's like turned into a real community.

Matthew: 54:19

And to hear the the the soy jack the soy teen, as they would call him, who did this, he had gotten onto 4chan because he was mad about the persecution of the soyjacks on the site, and then kind of sat in there and bided his time and collected information, and then launched this attack. And we and one of the ways that they signaled or they signaled that the the hack had been happened was they resurrected the qua board and then posted like a big soyjack message at the top saying that soyjack had won and posted dancing Hatsune Miku, like the the anime girl, Hatsune, can't remember her last name, dancing. It's like an animated GIF. And, of course, now it it appears that so this is all yeah. This is all like a big meme fight that's been brewing for literally years.

Jason: 55:16

So really, really, really painful series of sentences.

Matthew: 55:20

I know. I'm so sorry. I know it's one of those things where, like, you I start talking about it, I feel

Emanuel: 55:26

like It's okay.

Jason: 55:26

It's okay. Thank you thank you for saying them. It's just, you know, I feel that for for everyone everyone involved.

Matthew: 55:34

Because it's this thing that only, like, a very small group of young men probably care about. Right? These, like, fervent weirdo posters who care deeply about a five year old meme. I mean, again, they're gonna they're gonna go after me now.

Joseph: 55:52

And I and I think the last thing I'll just say, because sort of to hammer on hammer home the significance is that, yes, it is the combination of of a stupid five year old beef, but the hack and that information, like, that could have implications for law enforcement and

Emanuel: 56:10

Yep.

Joseph: 56:10

You know, federal agencies and, like, real stuff. So, yeah, even though the the context of it is pretty stupid, it could actually lead to some other stuff as well for sure.

Matthew: 56:22

And the I don't know if it's ever been down this long. Because it's been almost I think twenty hours that it's been just down, inaccessible. The thinking is that they they kind of in a scramble pulled it back down pulled it down because they realized how How old they were. Yeah. How how exposed they were and how brittle their back end was, and they're gonna have to rebuild everything.

Matthew: 56:45

And I don't know how long that will take or what that would look like.

Joseph: 56:50

Yeah. It might be well. Jason, any final thoughts?

Jason: 56:54

I was just gonna say that, you you know, like, 4chan has been an not up and down over the years, but it where it's located and who's running it and who's in charge of it has there's, a lot of twists and turns there. And Mhmm. I mean, I'm curious what happens now. I guess we'll see. But who knows whether this is existential, whether it's coming back soon, whether everything is gonna be wiped.

Jason: 57:19

Like, I don't know. But it's interesting. It's like it it it definitely is an important hack as you've all mentioned.

Joseph: 57:29

Yeah. And I guess we see what happens. I'm sure researchers are digging through the data. I'm sure we'll have another look as well. And if we do, we'll, of course, update the site and the podcast as well.

Joseph: 57:45

And just to take us out, I'll say, Matthew, we're all very happy. You're gonna be writing more regularly for the site. You've already done great stuff on AI and warfare and policy. And, yeah, thank you to our paying subscribers once again. I said this in the behind the blog, but thank you as the paying subscribers who can allow us to grow responsibly and sustainably like this because, you know, that's the entire point of four zero four media.

Joseph: 58:11

You know? Alright. With that, I will 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.

Joseph: 58:31

You'll get unlimited access to our articles and an ad free version of this podcast. You'll also get to listen to the subscribers only section where we talk about a bonus story each week. This podcast is made in partnership with Kaleidoscope. Another way to support us is by leaving a five star rating and review for the podcast. That stuff really does help us out.

Joseph: 58:50

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