The 404 Media Podcast (Premium Feed)

Hello, and welcome to the four zero four Media Podcast where we bring you unparalleled access in the

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.

I'm your host, Joseph. And with me are four zero four media cofounders, Sam Cole Yo. And Jason Kebler.

Hey. What's up?

So first of all, we are launching a new project, Jason. Do you briefly just wanna explain what that is about book bans, that sort of thing?

Yeah. So it's very early days, but we've partnered with our friends at MuckRock who is this nonprofit that helps us file FOIAs and helps other people file FOIAs, so public records requests. And we're gonna do a project where we are going to file a lot of FOIAs about book bans and educational censorship and curriculum changes and things of that nature, which has been something that we and many, many other people who you will be hearing from as we do this project have been fighting against since really, like, 2017, 2018 during Trump's first term. But we're gonna hopefully be documenting a bit about how this movement started from, like, trying to get teachers fired, trying to get libraries defunded, to taking over school boards, to then sort of, like, creating political power on the right through things like CRT, you know, railing against drag story time, things like this. This political movement has gotten very, very powerful.

And, again, lots of people have been reporting on this for quite some time, but we've wanted to do a project about this for a while. And so we were talking to MuckRock, and we're we're getting going on it. We bring it up because this is gonna be a long term project. So if you are a school board member, if you are a librarian, if you are a concerned parent, if you know something going on in your community, please tell us. Already, I think maybe a 100 people have reached out to say, like, this is happening in my town, this is happening in my school, this is happening in my library.

We also, you know, have done a lot of our own research about where to file FOIAs and that sort of thing. But if you have any ideas, let us know. And if you have already reached out to me, we've gotten about a 100 people who have reached out and we're going through that. I'll get back to you soon. But it's a very exciting time because we're gonna be working with Claire Woodcock who has written for us before, used to write for us at Motherboard and has been following this topic for a really long time.

So you'll start to see stories about that probably in the next month or so. The other quick bit of housekeeping is we have restocked our merch store. We have new designs. We have this sweatshirt is back in stock. Beanies are back in stock.

New t shirt, things like this. So go check it out. They're gonna be arriving in probably the next week, and it is far easier for me if you order now because I ship them all at once. So if you want that, go check out our Shopify page. You can find it at 40media.c0 and then click merch or store.

I forget which one it is, but you'll find I think it's merch. Yeah.

And I also googled four zero four media shop and it was the first result. So

worst what I do. That's how I find it. So peek behind the curtain there.

Yeah. Alright. Let's move to this week's stories. This first one is one that you wrote, Jason. The headline is meta tells workers building metaverse to use AI to go five x faster.

So the metaverse, it still exists?

Yeah. Allegedly, allegedly, it still exists. There's actually been like, I haven't spent time in Horizon Worlds or any of the other virtual reality prod products of meta for quite some time. Sam, I feel like you are always checking it out. Like, have you been have you been in the metaverse lately?

Have you have you strapped in?

I the metaverse always was, always is, always will be. That's that's the state of the metaverse. Mean, I haven't been in Horizon Worlds in a minute either, so I'm also shocked to hear that. I thought they laid off everybody who was working on it and, like, decided it was a bad idea, but I guess they did not.

Well, they have spent tens of billions of dollars on it, which is perhaps part of why this is this message is coming now. But the, you know, Meta, the company that renamed itself because of the metaverse, but formerly known as Facebook and now currently just obsessed with AI, like every other tech company, has decided to pay a little bit of attention to, its metaverse employees through the lens of use AI to to go a lot faster, to become more efficient, to do your job more and better.

And who who wrote this message then? Is it it's not coming from Zuckerberg. It's coming from somebody else. Right?

No. It's coming from Vishal Shah who is the head of well, their VP of Metaverse. So basically, he's in charge of metaverse development at the moment. I saw I did an interview with The Verge last week and, you know, it was fine, but I also tried to figure out what was going on with the metaverse. Like, it was an interview with him nominally about, like, developments in the metaverse, and it was so vague that I I could not tell what the current state of the metaverse is.

I have obviously talked to employees in metaverse development and just, like, in on that team, which is called Reality Labs. And so Reality Labs also encompasses the smart glasses that we've talked about a lot. So that is kind of like the shiny product in that world at the moment is the smart glasses. But anyways, Vishal Shah went on to an internal message board, which looks just like Facebook. Like, it's just a it's essentially like a Facebook group of four Facebook employees.

And the title is metaverse AI four p, which doesn't sound like it sounds like anything, but AI four p is AI for productivity. So the message says metaverse AI four p, think five x not 5%. Our goal is simple yet audacious, make AI a habit, not a novelty. This means prioritizing training and adoption for everyone so that using AI becomes second nature, just like any other tool we we rely on. It also means integrating AI into every major code base and workflow.

And then he goes on to say, I wanna see us go five x faster by eliminating the frictions that slow us down and five x faster to get to how our products feel much more quickly. Imagine a world where anyone can rapidly prototype an idea and feedback loops are measured in hours, not weeks. That's the future we're building.

So you see this message. And obviously, you read this and then maybe you see some reactions from Mesher employees or not and we'll get to that. What's your immediate reaction to seeing that message asking for five x productivity?

Yeah. So I mean, weirdly, I'm not surprised at all. And I got comment from Facebook or from Meta on this and they were like, I don't know why anyone would be surprised about this. Like, this is what we've been talking about for quite some time. And that's true.

Like, Mark Zuckerberg has been talking about trying to deploy coding agents all throughout Meta, the company. So basically, like, using AI to write code not just to help developers and programmers write code, but to just, like, deploy them and have them write and push the code themselves, presumably after some sort of human review, although maybe not always. And so this is like the future that Meta has been pushing for. It's the future that a lot of Silicon Valley companies and startups have been pushing for. But I think the reason that it's, like, very noteworthy is that it shows, like, the level of, I guess, delusion that we're under.

Like, it's it shows the level shows a lot about, like, how Meta thinks about its employees and how it thinks about its future. It's like, right now is a very difficult job market for computer programmers and computer scientists in part because a lot of Silicon Valley companies are not hiring because they think, oh, we'll just offload this work to AI. And this shows that Meta isn't like, oh, we can get a little bit more productivity from some of our employees if we leverage AI in a careful thoughtful way. It's like, no. Do five times the amount of work that you're doing right now.

So, I mean, think about that. It's like, do five days worth of work in one day. So I guess all of them are gonna be working Mondays and not any other day of the week, and everyone will be happy and and healthy, and that's how this is gonna work.

But even even if Meta deploys all of these coding agents in various parts of these companies or just in the metaverse to keep it to this example, how can you possibly five x a person's output? And then, of course, it's not even just coding as well. There's gonna be decisions have to be made inside meetings from human to human, just like it seems less moonshot, more delusional. Yeah.

Yeah. I mean, the the message sort of goes on to say, like, we have all these internal guides about how you can use AI and what you can use it for. You know, I don't have access to those guides. They're also doing some, like, events and webinars where they're like, here, use our coding agents. Like, use our AI tools to to become more efficient, blah blah blah.

But I think this really just shows, like, AI as being a tool to replace human labor even if that even if it is not, like, doing as good of a job as humans are doing. And that it's not like, oh, maybe we will just, like, close some job openings and we'll slow down hiring and maybe if we do layoffs, we won't replace people. It's like, no. The the future that they want is they want each human to be five times more efficient, to do five times more work. And, you know, I don't think that they're going to get there.

Like, lot of companies that have tried like, have replaced workers with AI have found that they actually need to hire human workers back. A lot of studies have sort of been done where it shows, like, AI can be helpful around the edges for specific tasks, but in many cases, it actually slows down work because you have to do a lot of, you know, like, quality control and fixing the fuck ups and hallucinations and things like that. But, like, this is this is like the job market that Meta is, like, hoping and imagining is one where for every human, they're getting five times the amount of work that they used to get, and, you know, you can do the math from there. Like, this portends layoffs. It portends, like, you know, hiring freezes, so on and so forth.

And I I think that is why it was notable and that's why I wanted to write about it is because it sort of shows how this, like, massively important company that I think still employs probably more than a 100,000 people worldwide is is thinking about AI and is looking at human labor.

Yeah. Even if you personally don't use a ton of meta products, or maybe just use Instagram, you definitely don't use Facebook or whatever, they're still a huge company, and they're still doing an absolute ton of stuff. So, you know, decisions like these are gonna impact a bunch of people's lives and then the users of the products as well. And even those who may not even be on Facebook, you you've still seen Facebook's AI slob even if you're not on Facebook. So did you see any of the responses from Meta workers about this or get a vibe of that or Yeah.

They're really I did. I did. I think the t like, Meta's huge company. There's a lot of people working on the metaverse. It's it was tricky in this case to include worker feedback because there's only so many people who, like, would have seen this message and kind of, like, sharing different things would have made it easier to figure out, like, basically, like, when a specific worker saw this or or whatever.

Like, so in that case, I didn't include much in the article, but but, like, vibes are low. The vibes are not good on the metaverse team in particular. I think that they feel a little bit like an afterthought at this point after, you know, so much money was poured into it. I think that, you know, Zuckerberg has shown himself to be interested in chasing whatever shiny new thing there is. In this case, it's AI and it's like, I don't know, they cloned CapCut, for example.

It's like, I feel like Meta in general is like constantly releasing products that are similar to other products that already exist and then they're like, wow. Here's our new thing and then they move on from it like really quickly. Like, I don't know if you guys remember, but they launched I was scrolling through my phone and they launched an app called Edits, which is literally a CapCut clone. They did that

heard of it.

And they did that because CapCut and TikTok got banned for a moment for, like, literally twenty four hours. And that's, like, when it came out and they're, oh, we are going to compete with Cap Cut. And I opened it the other day and it sucks. It's like it's not it's clearly, like, kind of being abandoned more or less. I mean, surely there's people still working on it, but I think that's how people working on the metaverse feel where it's like shiny new toys, Zuckerberg moves on, and then they're sort of still toiling away.

And then also, of course, being asked to do five x work, they're just like, fuck this.

Yeah. And then at the same time, Mark Zuckerberg is going with Palmer Lucky and announcing a new AR headset or whatever for The US American war fighter. And the poor people working on the metaverse is just like, well, do we have legs or not to these these avatars? Like, must suck.

Well, there there's that. And I think also there is honestly a sense I'm extrapolating out a little bit, but based on, like, the workers I've talked to and the stuff that I've seen, it's like there was this idea that they were gonna build the metaverse and it was gonna be a place where you go, like, surfing. Like, Zuckerberg is always showing, like, surf demos and, like, you know, concerts where you could buy little outfits for your avatar and and add legs and all of this sort of thing, like these developments. And, like, what it turns out that the company is interested in the metaverse for at the moment is, as you mentioned, like, military industrial complex applications. And so there is the, like, oh, we thought we were doing this to, like, connect people and to, you know, for, like, peaceful purposes, for lack of a better term.

And now it's like, well, the cops are using the smart glasses. ICE is using the smart glasses, and also we're, like, partnering with Palmer Lucky. It's, like, not not good vibe.

Yeah. That makes sense. And then I think just lastly, what and you touched on this with some of the the startups and stuff as well, but what what context is this message from Meta landing in? As in other companies are basically doing the same thing. They haven't said five x specifically necessarily, but, you know, Amazon is mentioned in the piece as well.

Right?

Yeah. That so there's actually been, like, a handful of companies that have been, like, messages emails from CEOs and messages from CEOs being like, I use ChatGPT and it wrote an email for me and it took notes at a meeting for me, and therefore, I expect everyone to use AI. And therefore, I expect you all to be, like, a lot more productive. And these messages have leaked, like, from time to time, and it's very funny because almost all of them have come from, like, quite small startups where they're just, like you know, you can kinda like roll your eyes and be like, oh, this person doesn't really know what they're talking about or like this CEO just found out about ChatGPT for the first time ever. And in this case, it's like, no.

It's like it's Meta. It's like one of the five largest companies on planet Earth, and that's what and they're telling their employees to do the same thing. And then similarly, Amazon CEO, Andy Jassy, said something like quite similar earlier this it was earlier this year where they said that they, quote, expect that this will reduce our total corporate workforce as we get efficiency gains from using AI extensively across the company. And so, again, yeah, most companies that are using this are sort of doing it in the name of increased productivity, but also, like and because of that increased productivity, some of you will not have jobs.

Great. Yeah.

And I mean, I don't it it's very hard to even tell sort of, like, the impact of AI on the job market at the moment because the Department of Labor and the Bureau of Labor Statistics and and things like that, like, are not releasing jobs reports. One, because when there was a bad one, Trump got very mad and fired the people who did it and then replaced them. But beyond that, the government is shut down now. So there there's not even, like, the kind of I mean, I'll say it, but like sham numbers or or the numbers that are like, you know, approved by the admin for release. And so it's kind of hard to say what the job market is doing right now, but like private estimates of how the job market is doing.

There's been, like, a few bank reports and, like, you know, like, consulting firms have done reports and they're, like, it's extremely bad. Like, very few places are hiring. A lot of places are doing layoffs. And most growth that we're seeing is in, like I could be wrong, but I think it's, like, nursing and AI. And it's, like, data center building, and it's, like, health care, and that's pretty much it.

And the entire rest of the economy is, like, very stagnant at the moment. So Yeah. Not good. Not good.

No. Not good at all. Alright. Let's leave that there. When we come back after the break, we're gonna talk about this really catastrophic Discord hack.

We'll be right back after this.

Okay. We're back. Gonna talk to both Sam and Joe about this. The title is the Discord hack is every user's worst nightmare. I feel like about a week ago, there was kind of, like, news about a Discord hack, but there was not a lot of details on what was happening, who was hacked, what was hacked, how how, like, the like, I don't know.

There's just, like, news about it, but then there wasn't a lot of details about what was in it. So what was in this breach? Like, what was the hack? What's the general background of what happened?

Yeah. Because I remember Sam actually saw something about this breach before we covered it, and, yeah, there were rumblings about it on Twitter from various info sec accounts. I think maybe Matthew, our regular contributor, saw something as well. But to peel back the curtain a little bit, it's kind of hard. You you can't just write a story.

Like, you have to find a way in, basically. And it's like, oh, people saying there might be a Discord breach. It's like, that's not that's not really an article. Fortunate for us, very unfortunate for the the victims of the breach and, of course, Discord as well, the hackers did start to publish some of this data that was related to Discord users. And, I mean, it's really, really sensitive stuff.

There are selfies in there, people's identity documents, and often those are the same photo because, of course, you have to take a photo of your face and hold up your ID so Discord can see or the third party service or whoever can see, yes, that's the same person and, yes, they're of a certain age or whatever, but there's also users' email addresses in there and their phone numbers. Sometimes this is linked to their customer support tickets. It's a bunch of really sensitive Discord user information. I mean, we'll get into it. It's not a breach of Discord.

It's not like Discord itself was hacked. It was a third party provider, but all of the data is Discord related. So I'm just calling it Discord breach, and I and I think that's totally fair. But, yeah, this is, like, really, really scary bad stuff that's so much worse than, like, I don't know, a username and password. This is, like, more sensitive.

Yeah. Is the company that was hacked, in this case, the company specifically dealing with age verification for Discord, or is it or was it more than just that?

So Discord has been a little bit flippy floppy on this information. They say that the ID the IDs taken were related to its age related appeals. As we'll get into, that's not necessarily the same as age verification, which Sam could talk about a lot, but it still shows it's basically the same risk. Right?

Got it.

Yeah. And then and I didn't put this in the article because it came out maybe maybe a day or two later, but Discord did tell me that the third party service provider was a company called five c a, which I'd never heard of before, but they were the ones that were actually hacked. Yeah.

Yeah. Okay. So hold that there. We'll get back into that in a moment. But so tell me a little bit about, like, how this information came out.

Like, you sort of watched the data coming out more or less in real time. Right?

Yeah. So Discord was being extorted for a while. As we said, rumblings of this breach had happened and were going around. I then got a link to this Telegram channel where the Discord hackers were taunting Discord, taunting, I think, Zendesk as well, which is, you know, a piece of software and a platform that apparently this third party probably used. And they were very annoyed because apparently Discord was not playing ball with this extortion attempt as, you know, readers will know, and you've probably seen because there's a million reports on it reports on it now, is so so common today for hackers to break into a company, steal the data, and they won't just send like a private extortion demands.

They will do this like spectacle, and they may even send information to journalists. They may send a very well choreographed or orchestrated email laying out to journalists, hey. Here's what we have. Blah blah blah. Or they may post it to Twitter.

That doesn't happen so much now because Twitter's changed its policies. And in this case, they were posting stuff to, Telegram. So I joined the channel. The hackers, as I said, are annoyed, and then they just start posting some of this user data. Now we only included heavily redacted copies in the article itself.

Like, we basically blurred out the entire face, hair, eyes, hand, username, ID of some of these people, and they were very, very small snippets. So it's just like a a handful of photos of, you know, what looks like a young person, or they are definitely young people holding their IDs. But then there were screenshots of what the hackers were presenting as much more stolen data. And I was just sat there in real time. I just joined the the group, and this very steady stream of data just keeps getting posted to the channel.

It was kind of a unfortunate matter of being, you know, in the right place at the right time. And, you know, this is just the way that some of these hackers do it nowadays. Yeah.

Yeah. Okay. I wanna pivot to the age verification slash age appeals situation, like, part of this. Because, Sam, you have been writing for a year, two years at this point about age verification laws and about sort of, like, what happens when users are required to provide an ID to access different web services. And in most cases, it seems like a lot of companies are outsourcing this work.

Like, they're having a third party company do this sort of thing. So age verification law or not, like, there are selfies of people holding up government IDs and their face, and that's, like, what was in this breach. This seems like the nightmare that you have been writing about from a theoretical perspective for a while, but now it has happened.

Yeah. I mean, this is what everyone's been warning about who's been talking about atrial defecation and the risks that go along with it. I was just looking it up because I was like, has it only been a year? And Louisiana passed their age representation law, and they were the first one in The States to really enact one of these laws in 2022. So it's been three years now of two and a half, three years of talking about the risks involved in these laws, and now I think we're finally seeing it happen in real life.

It's no longer a hypothetical, oh, maybe hackers will get access to this information. It's like, okay. Now it's become widespread enough. Enough platforms have it that bad actors are savvy to how it works and know that there's basically a honeypot waiting on any of these platforms that have a trification in place. And I think this is I mean, it's pretty it's a pretty basic tenant of, like, online privacy and security that the more places you have your stuff spread out there, the more of a risk it is to have your data breached or hacked or leaked or accessed by people you don't want it to.

And now you have to verify your age to get into certain Discord channels. You have to verify your age to get into Pornhub in a lot of different states, or you can access Pornhub in a lot of states, period, because they know these are the risks. But in the states that you do, you have to put your ID in. There's states where you have to do this for Blue Sky. It's just it's becoming more and more common to hand over not just your ID, but also, like, take a selfie with your face.

Even, like, tea, when we wrote about that hack, that was a pretty clear example of the way that verification systems that rely on government ID and selfies and these kinda know your customer systems that are these third party systems are vulnerable to this sort of thing. And I think it's it's sad, and it sucks in a couple ways. I think one of the ways that I think is the most striking to me is how this is gonna erode people's trust in pretty much any kind of security and privacy online in general. I think already people are like, I don't give a shit. Just tweet my Social Security number, it doesn't matter.

Like, no one cares anymore. And now it's like you're facing a prompt that says, we store your data safely. We don't store your data. We store it for three days, whatever it is. We make sure that we handle that correctly.

And then two weeks later, your selfie is on the Internet because hackers have found it, and you're holding up the picture of your ID. Yeah. It's like, at what point do people just, like, not care anymore what what a platform says about their privacy practices and just say, give it to whoever. I don't it doesn't matter to me anymore because I don't think the reaction is gonna be, I'm gonna lock down my op sec.

Or, like, I'm not gonna use Discord or I'm not gonna use like, my all my friends are in Discord. All the people I play video games with are on Discord. I'm just gonna not go on there. It's like that that's really not not that many people are going to just say, okay. I'm not gonna use these services anymore.

Like, a lot of people are just gonna do what they're being told to do.

Yeah. And we saw it with tea. It's like tea got more popular after the hack because it was in the news more. You know? It's like, I think I think we're just in such a, like, a nihilistic zone with our privacy that that is more worrisome to me even than, like, your data being out there.

It's like the the effect on the way we see our privacy is eroding.

Yeah. I mean, one thing that, you know, you you touched on it, but it's like, as a user, you cannot choose what third party age verification service, like, any of these platforms are going to use. And so there's a lot of different ones. Some of them say that, like, okay. We take your selfie.

We verify it, and then we, like, destroy it immediately or we encrypt it or we keep it for three days, then we get rid of it or whatever. And so you're kind of banking on the security of that third party, you know, age verification service. But if and not all not all companies do that, and, like, it's very hard as a user to vet the security of any given service that does this. And as these laws get more and more onerous and as they're in more and more places and as the definition of what porn is gets bigger and bigger or what adult content is gets bigger and bigger or it's like I feel like it's becoming very, very difficult to avoid this. So, yeah, maybe like one service does take privacy seriously, but another one will contract with someone that that doesn't.

I mean, even before this, we had something of like a halfway between a hypothetical and a real situation when a security researcher found that a ID verification surface for TikTok, UberX exposed people's driver's licenses. Like, there was some sort of exposed database and as part of that flow, the research was able to get those licenses. And in that case, that was a security researcher trying to warn the company to get the issue fixed and that sort of thing. No evidence it was, you know, done maliciously or used by hackers or entered by hackers. Well, now the full on real one has happened, which is like, these are hackers trying to extort people.

Well, extort Discord. Yeah.

Can I read you something unrelated that I just saw that is sort of related, which is Sam Altman just tweeted maybe the craziest tweet I've ever seen, which is, we made ChatGPT pretty restrictive to make sure we were being careful with mental health issues? We realized this made it less useful slash enjoyable to many users who had no mental health problems, but given the seriousness of this issue, we wanted to make this right. Then they talk a little bit about how they believe they have solved the mental health issues associated with chat GPT, which, okay, huge, like, claim to make. Goes on to say, in December, as we roll out age gating more fully and as part of our treat adult users like adults principle, we will allow even more, like erotica for verified adults.

Let's go.

So ChatGPT is getting into the porn chatbot business seemingly, but behind an age gate. And, like, how will that age gate work? I don't know. May uploading selfie, like, unclear at the moment, but I don't know. I I I literally just saw that as as we were reading it and the idea that they're that they, one, have solved mental health on chat GPT while they're being sued for, like, various people who have died because of advice that chat GPT gave them and also we're gonna roll out porn but behind an age gate.

This this tweet is, like, micro targeted to Sam because it has the mental health AI chatbot stuff in there, and we're now gonna give porn to adults. Also behind an age verification gate, Sam, what do you think of it?

It's Sam versus Sam. It's Yes. He's reading my my browser history and is like, how can I piss her off? Yeah. Yeah.

He's reading four zero four and concocting schemes to for us to blog about at this point.

Yeah. So, I mean, that was an interruption, but it is this is, like, front and center at the moment across, like, I feel almost every service that we use online because I think that one, like, as as these age verification laws are being pushed more heavily, you know, a lot of services have to put it in. And then the other thing that's happening is services are voluntarily age gating things because they're being blamed for bad outcomes happening to children because they've, like, rolled back all of these content moderation things. And so they're like, oh, like, anything goes as long as you prove that you're an adult and prove that you're an adult by, like, giving us your ID and a selfie and so on and so forth. There was one, like, little, not a quibble, but let's clarify the difference between an age related appeals process and age verification.

I think here, it's like what happened shows the danger of age verification, but age related appeals is perhaps something slightly different. Joe, you got in the weeds on this. Right?

A little bit. Yeah. As far as I understand, age related appeals is not the same as age verification. Like, you go to a help page on the Discord website and it says, help. I'm old enough to use Discord in my country, but I get locked out.

That's when you have to, you know, appeal and provide your ID. And, of course, in lots of circumstances, you know, and especially the social media network or something like that, companies are not supposed to provide products to say under thirteens or or something like that. So that would be more the age related appeals. And then to me, the age verification stuff would be like the law in The UK or the stuff that Sam was talking about, like these particular

The state laws and The state laws.

Yeah. Yeah. Yeah. And I feel like we're still we're still obviously seeing the rollout of those. So, I mean, there's no indication that the breach necessarily impacts age verification stuff.

The the two driver's license or the two IDs we saw, one was Ontario, Canada, and the other was, I think, Colorado, obviously, in The US. So it is impacting stuff in North America. It was 70,000 IDs in total, which doesn't sound like a lot in the grand scheme of data breaches. But, that's really, really sensitive stuff, and it could be linked to their other activity. And I guess just to clarify, that last thing I said at the top that I didn't put in the article just you know, the day later or whatever.

But, yeah, five CA was the the third party service provider, the Discord. It blamed it on them. And I'll just say, it's pretty unusual. Like, usually, a company that is some somehow impacted by a breach like Discord or whatever would just put out, hey. We're impacted.

It was a third party service provider, especially in the case of, those Snowflake breaches a while ago. Right? Not many, as far as I can remember, of the blog posts or the company said, yeah. It was Snowflake and we didn't have 2FA and that's why we got screwed over. Pretty unusual for Discord to say this, but maybe it was because the the hackers were kind of throwing Zendesk after the bus under the bus in in their telegram messages.

So I don't know. Just a a weird wrinkle there. Alright. Let's 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 four zero four media subscriber, we're gonna talk about what happened when AI came for craft beer. I mean, it wasn't great. You can subscribe and gain access to that content at 404media.co. We'll be right back after this. Alright.

And we are back in the subscribers only section. I've I've gotta just like not write the Google Doc just to make turn into fights. I should be nicer in in in these questions.

Joseph has written, Jason likes to drink water slash apple juice, question mark.

Referring to your taste in beer. It's like, I don't know.

The waterier, the better because it is healthier for you. And by healthier, I mean, you can have more than one or two Right. Without feeling very drunk and ill.

Having eight Miller Lite's, I don't know, is healthier than having one double IPA.

And this isn't a hypothetical. This is literally

I'm with Jason on this, just FYI. Yeah. I'm I'm pro watery lagers.

I'm pro Pacifico, Modelo, and Miller Lite, Coors Light. And then, like, under under other circumstances, like, I'll have lagers and pilsners, but these, like, disgusting double IPAs, coffee, stout, etcetera, etcetera that Joseph drinks. And I believe Emmanuel as well. They have a little beer club.

I will speak for Emmanuel even though he's not here, but I'm sure I can speak for him because I know enough about his beer taste and, like, yeah, that's absolutely right. I'm really into, I don't know, 7%. 8% is like borderline, but like a 7.5 is perfect for a double IPA or a stout or a porter. I mean, really, any any beer, really. And, yeah, I got Emmanuel into it, and he has his own taste as well, but pretty similar.

I love it. Yeah. We trade our little tasting notes. It's not I I don't know what I'm talking about. I sniff it, and I'm like,

I don't know.

Smells light, I guess. Tastes good. It's strong. I should do a I should do a class or something.

We should do a podcast about beer in general because with descriptions like that, I mean, it's just thrill it's thrilling stuff. So this story starts around April with a competition. What what happened? What what is the story, first of all?

Yeah. So headline, what happened when AI came for craft beer? Initially, this article was just gonna be about this thing that happened way back in April. And for various reasons that I'm sure subscribers are sick of me hearing about, but it turned out like ICE started doing a bunch of stuff. So this just fell to the wayside and was a shame because, you know, I think it's really interesting.

That said, I left I left the article on the back burner so much, the stuff actually developed and then it became a much more sort of conclusive complete story. So way back in April, there's some sort of beer competition from the Canadian Brewing Awards to be a judge there. You're a volunteer, but you have to pass an exam basically. I think it's a written one and then probably some tasty ones as well. The brewing process, different beer styles of that sort of thing.

And CBA, the Canadian Brewing Awards, I mean, they're rarely prominent. Tons of breweries will enter their beer for most best experimental beer of the year or whatever. And usually, what you do is the judges will go there, they'll taste the beers, they'll make the notes or whatever. And in the roundabout on the third day of the competition, I was told the judges were asked to enter their tasting notes into some sort of new AI powered app instead of the software platform they were already using. And this is according to Greg Loudon, who's one of the judges I spoke to on the record, and another judge who I gave anonymity to protect him from retaliation for reasons they'll probably make clear in a minute.

But they they start entering this stuff into this app, and it basically looks like a chat GPT wrapper. And from the description of one of these judges, it was pretty bad. It was quote, parroting back bad versions of your judge tasting notes. There wasn't really an opportunity for us to actually write our own evaluation. And it would then provide additional questions for the judges to answer that were, quote, total garbage.

So they weren't particularly happy about that. And then, you know, some of them, they they get together and they start talking amongst themselves, figuring out what to do, and then that's where we get to this open letter.

Yeah. So open letter from the judges because they're mad. I think not much more to say is than that. So the judges revolt, basically.

Yeah. And let me just open up the letter now, which is still online. But, yes, this was published April 5 by Greg, and then it's signed by I'm scrolling down. Well, I'm not gonna count if anybody else has added here, but it's something like 30 people signed onto it. It's like breweries and and then judges and that sort of thing.

And the letter says, quote, you know, they introduced this AI model to their pool of 40 plus judges in the middle of the competition, surprising everybody. And it's, of course, complaining about that, and the judges believe that their tasting notes were being used to train this AI. And I'll just be very clear, even though probably repeat this in a bit. The Canadian Brewing Awards slash Best Beer, which is the company affiliated with this, they say no. This data was not being used to train an AI.

They were not making their own model. Instead, they were using some third party AI for their purposes. I presume it's Giant GBT or or something like that. But the judges are very, very mad at the perception of that. And then even just on a more like fundamental level, they just don't like the idea of AI coming into beer tasting, brewing, and judging, which is, you know, a very human endeavor.

Greg, you know, later tells me that brewing is an art. You don't wanna bring AI into this thing and make it sterile and the same. And the the letter concludes with asking their fellow judges and industry owners and all of that, sign a letter, spread the word, raise awareness about the real human harms of AI in your spheres of influence, have frank discussions with your employers, colleagues, and friends about AI use in the industry and their lives, demand more transparency about competition organizations. I I mean, the reason I wanted to write this is because we've covered how AI has impacted illustrators, artists, voice actors, musicians, like all of these other industries. And like they're almost unfortunately, like predictable.

Like it's obvious that AI is gonna impact voice acting or music or whatever. It wasn't I don't think it was on everybody's bingo card that AI would impact professional pitch, I think. Did this this open

letter is crazy, by the way. Like, I say that complimentary, but it is a really wild open letter. I I read your article, but I didn't, like, read every bit of the open letter, and their concerns are numerous. Yes. One judge said, quote, I am here to judge beer, not to beta test.

They were also talking about how the addition of the AI meant that a lot of the breweries had to send a lot of extra beer for some reason. That part is not not well written in the open letter, but they say that, oh, a consumer focused evaluation. So basically, like, the they had to send extra beer because they were gonna have, like, non trained judges use this software, I believe. It says 3,200 extra cans of beer were required. To put it plainly, the introduction of generative AI significantly reduced the efficiency and speed of the competition.

It's it's weird because this is actually a pretty complicated story. And I guess that's why it also took me a minute in that the AI was introduced in the in this specific competition back in April or wherever. But also, Best Beer, the company behind the CBA, has been developing a consumer facing app. You know, like, untapped, basically, an app

to track killer according to the open letter.

Exactly. That was the intention. And to be perfectly honest with you, when I spoke to representatives of Bespear, it was difficult to get a straight answer out of them, basically. And I think that was another benefit of letting time pass a bit because, like, oh, now stuff is clearer. But at the time, they they just weren't being straight with me, but I managed to eventually determine, yeah, they were gonna make a consumer app as well.

They would use AI to recommend beers to people. So there's the judging side and the commercial app side. And I I I think you can see why the judges came to their conclusion that while our notes are being used to train the AI, and even if that was incorrect or appears to be incorrect, it was still thrust on them basically without consent. It's like, hey. Use this AI tool.

It's like, I don't know. I'm a volunteer who has probably drank beer my entire life and has expertise. You want me to use chat GPT wrapper? Like, no. I'm good.

You

know? The robot only one robot who knows that beer's bender from Futurama. Anyways, there's one word there's one word passage in the open letter that I think is really good. The use of AI is dehumanizing, stripping away all of the individuality in judging and palette variations across people at different environments, race, gender, and genetic makeup. There are already recorded instances of racial bias among generative AI models.

Anonymization and aggregation of data goes against the core tenets of beer judging, removing accountability from the equation and denying recourse against corrupted impaired bias or belligerent judges. Beautiful.

There there was too much to include, to be perfectly honest.

Oh, yeah. Of course.

Yeah. Yeah. Yeah. It's it's a long letter.

Okay. So did we already talk about Bess Beer's response or no?

No. Well, I touched on it, but I'll just say that when I was talking to them originally, I spoke to the head of innovation at CBA, they said, breweries will have amazing insight on previously unavailable useful details about their beer and their performance in our competition, and then craft beer drinkers will able to better sift through the noise and find beers perfect for their palate. That's what they told me at the time when they were making the app. Well, what happened later? And I'm skipping one of the questions forward a little bit, just because I'll say the response to this as well.

Well, CBA basically stepped away. Best beer stepped away, and it kind of fell apart, it seems. They did an Instagram post saying the team is is is not gonna be involved with this anymore, blah blah blah. And they said in Instagram, yes, they were stepping away, the goal was to make make medals matter more to consumers. And then in a email to me, they just said the best beer project was never designed to replace or profit from judges.

You know? They they took real issue of the open letter, I I I should say, and they did threaten Greg with, like, legal action because they felt it misrepresented their brand and what they were doing. I was also, you know, sent a bunch of those texts between Greg and Bespier, and there wasn't enough space to even get into it. But, yeah, there's been a lot of fallout from this sloppy miscommunication from CBA and forcing this AI on them, basically. Yeah.

Yeah. Okay. So time passes. Things died down until September. So what's the what's the conclusion here?

Well, yeah, the the CBA says we're stepping away and, you know, I I reached out for comment and they they take a little while to respond and they say, oh, actually, we posted to Instagram and that's where they say, yep. The the team's no longer involved in this. But, know, the judges the judges I spoke to well, Greg, especially, he wanted an apology. And now that has apparently come from CBA in the Instagram post. I think he is blocked by CBA on in see, there's all of these, like, little these other tangents you can go to about the drama that I didn't really get into.

But then seriously and I didn't get into this either, but I think Greg has, like, left the organization and just doesn't wanna be affiliated anymore because it was, like, a lot facing all of this legal pressure and that sort of thing. So as an outsider, it might sound trivial or whatever, but I don't know. This is kind of fucking up people's lives a little bit, you know, in their hobby or sometimes their professional stuff as well. And I think I'll just leave it with what one judge says. They told me, I don't think anyone who is hell bent on using AI is gonna stop until it's no longer worth it for them to do so.

So I don't know. I'm pretty sure this won't be the last instance of someone trying to shove AI onto craft beer. And I should say, really interested in making a beer. So if there are any breweries, I don't I don't know how many of our pay subscribers work at breweries or something, but let us know. I think there could be room for something there.

Four zero four medium Mexican lager, make it happen, please.

That that's the one thing. Because if we're going to do if we're gonna collaborate with a brewery, which we would love to do, it would be very selfish of me and Emmanuel to just make a 7% of IPA called AI slop or whatever. It has to be one that you and Sam and anyone really would like to drink as well. So I think a 6% IPA. Okay.

Dude, come on. That's really

You know

was the beer I really liked? Old Blue Last.

And where does that come from exactly?

Satan's piss. I don't know. People can Google it.

Old Blue Last was Weiss's beer. Short lived. Not good. Tasted like metal.

I only had it warm because it came from, like, someone's underneath their desk.

It was never ever ever ever in a refrigerated environment.

I don't know if it's the taste. The flavor profile came out that way.

Yeah. Good good god.

Good idea, Joe.

Sound well, it tastes like metal because that's how lager tastes. So I think that's just how it was. Anyway, I have the outro script. Alright. And with that, I will play us out if we can find it.

Here we go. 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. 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 helps us out. This has been four zero four Media. We will see you again next week.