Loading summary
Transcript81 lines
- [00:01]
Sam Cole
This episode is brought to you by Dutch Bros. Big smiles, rocking tunes and epic drinks. Dutch Bros. Is all about you choose from a variety of customizable handcrafted beverages like our Rebel Energy drinks, coffees, teas and more. Download the Dutch Bros app for a free medium drink. Plus find your nearest shop, order ahead and start earning rewards offer valid for new app users only. Free medium drink reward upon registration. 14 day expiration terms apply. See Dutchbros.com foreign.
- [00:34]
Jason Kebler
Hello and welcome to the 404 Media podcast where we're doing something a little bit different. I'm not even going to read the intro. We're going to talk about the previous year. We're not going to be talking about our stories that we've just published because this is coming out during the holidays. Hopefully you're all relaxing, chilling, and now you're going to listen to us reflect on the year, especially when it comes to AI and journalism and all of that. No subscribers section this week, but it's still going to be ad free for paying subscribers. We did want to give free listeners a preview of the sort of conversations that you can get access to when you do become a paying 404 media subscriber as well. So first of all, I'll just introduce everybody, of course, the other 404 Media co founders with me, the first being Sam Cole.
- [01:27]
Emmanuel
Yo.
- [01:28]
Jason Kebler
Jason Kebler.
- [01:29]
Unnamed Speaker
Oh, I get to go in the middle this time.
- [01:32]
Jason Kebler
And Emmanuel, baby, here I am.
- [01:35]
Unnamed Speaker
Dude, that you really switched it up. Come on. Like, yeah, I was not ready. I was not ready.
- [01:40]
Jason Kebler
Honestly, I'm a little tired and hungover. So that's.
- [01:44]
Unnamed Speaker
Oh, wow. Very interesting. So people know we're recording this, like from the past. You're listening to this in the future. So if anything, like, crazy happens, we're recording this December 16th. Just so you know.
- [01:59]
Emmanuel
I hope nothing crazy happens. I need to, like, not have anything chaotic happen at the end of the year. There was one year where like, pornhub changed its entire terms and like wiped its sight of all of its content at the end of the year. And that was my worst year yet.
- [02:13]
Jason Kebler
Like, you had to jump on and.
- [02:15]
Emmanuel
Yeah, it was like major breaking news. It was like we were going to break that news and like, it was like the very end of December. It's like, Jesus Christ. I was fully checked out.
- [02:23]
Unnamed Speaker
If the, if the singularity is achieved in the next week, that will not be incorporated into our talk about AI.
- [02:30]
Emmanuel
We're going to pretend we don't see that.
- [02:31]
Jason Kebler
Yeah, I think so. So right now the plan is that we'll be airing an episode about how we cover AI that was for paying subscribers a long time ago, basically, at the time we launched. So if things go to according to the schedule that we have right now, you would have already heard that, you would have already listened to it. But here we're going to talk a little bit more about AI and I think much more of the sort of current criticism around it and all of that sort of thing. Jason, do you want to start, like, how do we get into this in how people are covering AI now and sort of the current debate around that? I know you and Emmanuel have a lot to say about this.
- [03:19]
Unnamed Speaker
Yeah, I mean, we've talked about AI almost every episode of this podcast for the full year. And I think that we wanted to try to say, like, here's the current state of things, I guess. I think that we have tried to focus on how AI is actually being used right now. Um, we've done a few articles about sort of, like, where things are going and the actual, like, technology and, and how that all works. But I think that we've been quite interested not in covering every update to, you know, OpenAI's large language models or, you know, different models that have been released or what their use cases are, but. But covering how they filter down into the real world and how they're used, and then also focusing on, like, where the data is coming from, like, governance of them and that sort of thing. But there's been like, a big debate about whether generative AI is useful at all. Sort of, like, what is going to happen to it as they run out of data to train on. Like, Ilya Sutskever, who is one of the founders of OpenAI, recently said at NeurIPS, which is the industry's biggest conference, that they're pretty much out of data. They're out of human data to train on. And there's different schools of thought as to what happens after the data, the data is gone. The large language models are going to have to be trained in different ways. The AI is going to have to be trained in different ways. The leaps forward might not get as good. You know, you might have AI training itself on other AI and having, like, this recursive situation where, you know, AI models are trained on AI slop, more or less. And that has been like, a big topic of conversation. I think we'll probably talk about it a fair bit in this podcast. But there has been a lot of discussion about a column that Casey Newton wrote about about sort of like whether.
- [05:43]
Jason Kebler
AI is of Platformer, which is Platformer.
- [05:46]
Unnamed Speaker
Yeah.
- [05:46]
Jason Kebler
Independent publication.
- [05:47]
Unnamed Speaker
Yeah, yeah. So I don't know, do we want to talk about that debate now or do we want to talk more about how we cover AI and what changed this year and then get into it?
- [05:59]
Jason Kebler
I think we can go into the debate. All I would say is that we, I think it's quite obvious to anybody who has read our website in that, well, arguably we cover. Well, we cover AI not just sceptically. I just think realistically in the way that we're not. We don't go out and say, oh, AI is never going to take off, this is complete rubbish or anything like that. It's more, as you said at the top, it's much more about the harms and the impacts that happen on humans. Right now, whether that's all of our work on nudify apps or deep fakes or non consensual imagery, I haven't done anywhere to the same quality that you've all done done on that. But I'm trying to look at more military uses of UI and AI and I'm especially interested in that. The tone of our stuff is, I would say, more realistic than anything. But yeah, what is this column exactly and why has it triggered so much, I would say, conversation around how people are even framing AI.
- [07:11]
Unnamed Speaker
So Casey Newton, a platformer, longtime Verge reporter, started, you know, a newsletter called Platformer that initially covered content moderation on the Internet and has grown to prominence. Casey is like very well respected, has a podcast called Hard Fork at the New York Times. And so I think when he writes things, people pay attention to them. And last, well, earlier this month he did an article called the Phony Comforts of AI Skepticism. And the subhead of that was it's fun to say that artificial intelligence is fake and sucks, but evidence is mounting that it's real and dangerous. And I'm trying to think how to say this, but. But basically the argument that Casey made was that the people who have a knee jerk reaction to say that generative AI is not going to change everything are wrong in his view and that this is like a naive perspective, more or less. This has started like a massive conversation about the state of AI, I guess. And my personal opinion is that the way that we cover AI has not been really addressed in Casey's pieces, nor in the rebuttals to Casey's pieces, of which there's been like various back and forth and I don't think we're gonna. I haven't read every single back and forth Because I find it a little bit exhausting. There's like 20 different blogs about the blogs that go back and forth. But Kasey sort of argues that the people who are saying generative AI isn't going to change things because it's just not good enough are wrong and sort of says that people in the industry believe that it's going to change things. And then there are a lot of rebuttals to that saying, like, Casey, you don't know what you're talking about, blah, blah, blah, you've covered this. And then we can talk about where I come down on it. But, like, how would you frame the debate? I guess so.
- [09:25]
I would say that I don't think there's anything terribly wrong with Casey's column. He's just describing kind of the far ends of the spectrum, which in his framing is there's a group of people who think AI is like a giant hoax, much like, I don't know, the NFT boom and web three scams that seem very real and people in tech swore up and down are going to change the world and then have just like completely disappeared. And then on the other end of the spectrum are people who are like, you don't understand. We are inches away from general intelligence and this is going to change everything. And we need to take this very seriously because AI is all too real. And I think it is correct that that is what the two far ends of the spectrum are saying. But obviously it's a spectrum and there's a whole lot of variety in between. I think the best rebuttal to this that I've read is from our former colleague at motherboard, Edward Nguesso Jr. Who wrote on his substack and the gist of his argument is that Casey's taxonomy is bad, which is it's not the case that AI is either fake or real and dangerous. It can be all of those things. It could be a mix of those things at once. And that certainly is something that we've seen as true in our reporting rights. Like an AI technology can not work at all, but still be implemented and be dangerous. Right. If you think about some AI software that is very faulty but is actually being used to filter job applications, Right? That's a case of AI being quote, unquote fake, but it is actually being used and it is actually having real world impact and is dangerous. And the opposite combination of things can be true also. I think that's a really good argument. People should go read that Ed. To complicate the. The Take Factory Even more, Ed quotes a blog from this guy called, I'm probably mispronouncing his name, but Ali Al Khatib I think it is. And he says, quote, AI is an ideological project to shift authority and autonomy away from individual towards centralized structures of power. And I think the idea that AI has always been kind of a loose term that can mean a whole bunch of things, and to think of that more as a political project is a pretty useful idea, right? So again, to take like the job application example, you can kind of shift the responsibility of discriminatory hiring practices to an AI model that isn't really doing anything that fundamentally different than a bunch of automated systems that are already in place to filter through applications. But by calling it AI and putting all the power on this system, you're kind of perpetuating discriminatory hiring practices, but you've shifted the, the power and the blame and the responsibility to this AI system rather than like the HR department or the CEO or whatever. And that's something we see happen with AI a lot. So I thought that was a really good idea. And that blog, Ed's blog, and I think the whole discussion has been like, pretty useful. Like, definitely annoying to see a bunch of blogs about blogs, but I think it's like a necessary conversation. I would say that, having said all of that, two things. One is, I do think that our perspective is kind of outside of the debate because we don't normally have these type of super zoomed out conversations about AI and what it all means. We did have it at some point earlier. No, it was almost a year ago at this point. But it's like we did that because we realize that we didn't plan for it, but basically every story that we were writing was about AI. So we did take a step back to address that because that's not something that we historically have done as writers or as a publication at Motherboard. Right. It's like we've never been on one beat for that long. Usually things change and a technology fades or a new technology comes along and we all shift with it. But we were focusing on generative AI so much that we wanted to address it. And I think what we said then remains true. And it's kind of outside of this debate, which is like we're focused on what people are actually doing and that is what we continue to focus on going forward. And that just kind of a little bit removed from like this more zoom doc conversation about like, is AI going to destroy the world or not? I don't know. But what I do know is that it's fucking up Google search results, that it's being used to make non consensual porn, that it's fucking up Google books and all this stuff. And I think that is where our reporting is most useful and has the most impact. So that's where we remain. The other thing I want to say is when we had that conversation last year about AI, I kind of laid out the ideological battle that was taking place in AI, which is like, I think back then they were calling it like the AI safety ism versus the AI accelerationists.
- [15:47]
Right?
- [15:47]
Accelerationists. Right? Yeah. And the accelerationists have definitely won the debate. Like the Mark Andreessen and Elon Musk of the world are in power. They're controlling these companies, they're leading this technology. They've definitely won that conversation. And Marc Andreessen, who is one of the biggest figures in this debate, he put out this manifesto that he called the Techno Optimist manifesto. And all I would say to that is that I think it is weird that it's like there is the AI as powerful and dangerous perspective, and there's the AI as fake and a scam perspective. But I'm not hearing a whole lot of AI could be powerful and maybe it can be good. There's no one in the conversation that is like, what if technology but good? And I think when we zoom out and we talk about technology in terms of these big ideas, that is our perspective. That's what we have to keep explaining to people who say that we hate technology. It's not that we hate technology, it's that we focus on what it actually does. And when you focus on what you actually, what it actually does, you see a lot of bad examples. And that is what we want to shed light on because we want to have positive impact. But if you were to ask me, like, where do you see it all going? Or like, where do you want to see it going? Then it's like, there's some cool tech, it has some potential, and I would like it to be used for good rather than bad. And it's just like if there was a robot that was able to like do all my chores around the house and like take that burden off of me, that would be great. And the fact that there's just like no one in this conversation that is staking that position really is, I find to be just kind of a bummer. I don't know. So that's, that's just like the only thing that I would add to that like, whole debate is that the, like, the people who are optimistic about the future are like, Marc Andreessen. And there's no one from, like, I don't know, the Edward side of the conversation, which is saying, like, what if we had some kind of technology that was like, good for people rather than bad?
- [18:04]
Jason Kebler
Well, do you think there is nobody taking that position because there aren't those good use cases or because taking that position is just not worth it for them?
- [18:15]
Unnamed Speaker
I think it's two things. One is, since Marc Andreessen and co have staked that position, then, like, you're in opposition to them, so naturally you're like, well, he's full of shit. And it's like he's lying, which is certainly true a lot of the time. And then, I don't know, there isn't like a ton of good examples. Once again, right. Like, when we. When we look at what people actually use this for, it's like, it sucks. Like, a lot of the time, it just sucks. And that's not me being negative. That's just like me opening my eyes and like, looking what at what people do with AI image generators. It's not good, guys. It's like, you know, it's like, I've looked at the evidence. It's not good.
- [18:55]
Right? So that. That is the position that I feel like is not super represented in this debate, which is that, let's say, like, I think I talked about this in my TCL AI Generated shorts article a bit where let's say you imagine the greatest, like, use of AI in Hollywood and it, you know, improves special effects. It's used responsibly by people who care who. By people who know what they're doing. And it makes cool movies. Like AI is going to be used in Hollywood. It is already being used in Hollywood. It is being used probably in ways that are like, mostly fine, also ways that are not fine. But there are things that have been done using AI in big movies that it's like, cool. You couldn't do that three years ago, and now it's in a commercially released film.
- [19:44]
Jason Kebler
And people like it de aging Tom Hanks or aging Tom Hanks very quickly and much cheaper. I mean, that's. The example comes in New York Times articles talking about AI in Hollywood.
- [19:53]
Unnamed Speaker
Yeah, but then you look at like the. This tool is being released widely, and I think it has to be released widely. It's. There's no way of like, keeping these things from being used by anyone. And the ways that it's being used are to Spam, the Internet and the, and to make better, more convincing spam and everything that comes alongside of that non consensual pornography, disinformation, blah blah, blah like all of that stuff. And there are far fewer people who are using artificial intelligence in its like utopian, hey, let's use this for, to make cool stuff earnestly use like to augment human creative capabilities. And there's far more people using it to spam shitty books on Amazon, to spam Facebook, to spam, to like do targeted advertising and and so on and so forth. And then the actual like impact and result of that is that human made work is a lot harder to find. It's devalued. You have like, it's all this, all these like really terrible chatbots that make mistakes. And then you also have sort of the, the permission structure for all of these big companies to release their half baked AI and sort of like all of the shitty effects of that. And that is what we have seen in 2024 by and large. And I don't think that that is really even represented in, in these debates. It's like you have the utopian here is here's all the cool things that chat GPT can do. You can write a better email with it or whatever, like other stuff than that. But like it's going to save your employees for four seconds of time so that they're going to be able to automatically analyze meetings and stuff and therefore your company will get value out of them. And then there's like the fact that a small number or a large number of people can use it to create infinite pieces of shit that flood everything and like ruin the Internet. And that is neither like that doesn't fall into the shitty and fake side of things or the like real and good and gonna change the world side of things. It's like all of those things that as Edward sort of mentions with an emphasis on like things up.
- [22:22]
I feel it's also like it's the thing that's not discussed because it's like I find it very interesting. I find it the most interesting because it's real and it's bizarre. But it's like it's not this utopian dystopian thing. It's just like was it, is it shrimp Jesus that became like the word of the year or something like this slop, slop, AI slop. And it's just like that is like the accurate take because it's like that's the majority of the output. It's like we can have These big ideas about AGI and all that. But it's like when you look at what is actually happening, it's like overwhelmingly the result of this technology has just been slop, which is like an overused term at this point. But it's like, it's just true. It's just what is happening. It's like Casey Noon, Gary Marcus can like talk about all this stuff as much as they want, but it's like what is actually happening. It's just like ugly images on, on Facebook.
- [23:20]
Yeah. I have one more thing I want to say and then I want to hear what Sam thinks or just like I'll stop dominating convo. But when we were at Vice, like right before we left Vice, Nancy Dubuque, who was our CEO, asked me to make a presentation about AI for the board members, more or less. And then she was fired before I could give the presentation. And I pulled it up and the first thing, the first slide that I have is generative AI. And the fact that AI is not only generative AI, and I think the fact that Kasey does this in his piece, like, sort of mixes machine learning with like algorithms with AI, with generative AI, it's, it really muddies the water. It makes it very difficult to have a conversation about these things because we've been covering artificial intelligence for a decade. Artificial intelligence is here in many, many ways. It's changed the world in many, many ways. It's clearly, clearly real in many, many ways. And then the next slide, I have examples of that and I have social media algorithms, self driving cars, facial recognition, criminal justice algorithms, drug generation and disease diagnosis, labor robots, worker surveillance, customer service and language processing, smartphone cameras, hallucinating. Like if you take a picture of the moon, you're not really taking a picture of the moon, you're just like telling an algorithm to generate something there. Deep fakes, scams, malware, non consensual porn, high frequency trading, and like lots of other things. And these have changed the world in many, many, many ways. But when people talk about AI, it's all kind of just like lumped together. And I think that that is like super harmful. Like you kind of need to be specific about what you're talking about because AI and machine learning have changed the world in many, many ways that we've been reporting on for a long time. It's just a question of whether like this specific generative AI tool is going to lead to a super intelligence or like lead to everyone losing their jobs. And I think that that has been a little bit Lost in the debate. And Casey even says, like, I'm only going to talk about generative AI. And then he lists a bunch of stuff that is not generative AI. That's like more traditional machine learning and things like that. And that's been part of this whole thing.
- [25:53]
Emmanuel
But anyways, yeah, I mean, that's kind of like what you're talking about with, like, being told to make a presentation for an executive at a company that is scrambling to stay relevant or become profitable or dig itself out of a hole. In general, I think is kind of the. The heart of what's happening with AI as like a marketing term, because in the last few years, it has become a marketing term. It's this idea that's not even. It's like you said, it's not even. We're not even talking about machine learning. Machine vision versus actual what is AI? It's become a thing that you can slap onto products or slap onto your company and hope that it fixes whatever is wrong with what's going on. And I think in that way, it is a lot like the metaverse and NFTs and all this stuff that gets hyped. It has hype cycles. But because there's a separate thing happening, which is like, machine learning is changing the world, like you said, and you're using it all the time, every day, without knowing that you're interacting with like, quote, unquote, AI. You don't really know that you're using AI, whatever that is meant to mean, until you have one of those stupid little, like, sparkle emoji things come up in your face that says, like, would you like to rephrase that? Like, notepad entry? It's like, fuck off. And I think people have a strong reaction to being advertised to, and they know that they're being advertised to all the time in products that they already pay for. It's like, I don't need that feature in my Word document or anywhere or in my text field when I'm just like, writing a text. I think that's part of, like, where this has come from is these two things are happening at the same time where, like, executives are saying, we need to stay relevant, we need to make money, we need to strike these deals with these companies that are profiting off of our content or whatever it is. And at the same time, it's like users are getting hit with this garbage all the time, and they know it's garbage and they're sick of it. And those things coming up together becomes like a versus situation. And Then you have these conversations going on that are. I mean, I don't know, like, I don't want to speak too directly about, like, anyone's, like, journalism, but there is a real profit motive to starting a conversation that at its heart is controversial and kind of stupid. And, like, getting everyone to share your substack article probably is a good way to. I don't know how substack, like, rev share works, honestly, but it certainly gets people to subscribe if they relate to what you're saying about AI. Because AI has become this identity thing. It's like being against AI is an identity. Being for AI is an identity. And there's not a lot of in between that people talk about. But most people probably feel it's just like, I don't know, I don't have a strong opinion about it. I think it's something that's in the future is probably what a lot of people think about it.
- [29:01]
Unnamed Speaker
Yeah. This is another part of the presentation that I gave, because the entire point of the presentation was like, how do we VICE use AI to become relevant again and to, like, dig us ourselves out of this hole? And like, blah, blah, blah. And it was never like, hey, we're going to replace all the reporters. They just, like, ended up firing them anyway. But in that presentation, I had a whole slide that says, we're already using AI. It's like anyone who has a job anywhere in any sort of creative field is using artificial intelligence, whether they know it or not, as sort of like, table stakes to remain relevant in their job. And what I mean by that is like, the. The optimization of, like, VICE can be a better. Or like, any journalist can be a better, more productive journalist if they use AI has already occurred, and it's occurred, like, long ago. And the examples were descript. Like, we're using Riverside right now to record this. Riverside is a podcast tool that does a lot of the, like, editing process for us, sort of. It's like we have. It can. At least we're recording this with our podcast producer and editor, and she's made it, like, way, way better. It's kind of trash, but, like, if you're putting together a really, like, a narrative podcast and you need to transcribe everything and, like, cut selects and stuff, which is how narrative podcasts are made. You might do it in Riverside. Like, you might do a lot of that work in Riverside. You're not hand transcribing, which we all used to do. Otter. AI is another one that, like, a lot of journalists use, where you're like we, we used to just transcribe every single interview we did by hand. And I used to just type very fast when I was on the phone with people and I've stopped doing that and I now just like have a robot transcribe it for me and I listen back to it and make sure that everything is accurate. But it's like that saves me hours and hours and hours and hours as an editor. Advice. Reporters used to say, hey, I'm doing this feature. I'm just going to transcribe today and tomorrow and be like cool. Like you're just gonna type like all day. And it took forever. There's auto generated subtitles on like social media things and on YouTube transcripts, there's translation. Like Google Translate is really good now and it changes how we report.
- [31:36]
Emmanuel
But even with these like I, like someone corrected me on this recently and so I, and I didn't actually go further with it to find out, but I totally should have. But like I want to hear what you guys think. It's like those like is transcription AI even like, is that speech synthesis or is that AI? Like is, is speech to text actually generative AI is it. It's definitely like machine learning maybe. I mean like, like it's like where are these, where do these things land in this spectrum of like AI as a word? It's not creating something new, you know, I don't know.
- [32:10]
Unnamed Speaker
I know we just had a conversation about how this is important, but I think it doesn't matter that much, right? In terms of like when you're just like, yeah, I can save five minutes doing this and it doesn't matter.
- [32:20]
Emmanuel
But it's like robots is like the overarching kind of. It's like we've talked about robots a couple times. It's like, well, I get a robot to do that. Like people kind of consider that AI now, which is interesting. And AI has become this thing that people feel strongly about. But I use AI. I use AI the other day for. I use it all the time. But like I use it specifically to like find a shirt that I liked with Google lens. Like that's, that's, that's machine vision. I use it to where is it in AI?
- [32:46]
Jason Kebler
I use it to take photos of my food and automatically calculate the nutritional content.
- [32:52]
Unnamed Speaker
Right?
- [32:52]
That's not why. That's good at data.
- [32:54]
Jason Kebler
I know I've been playing.
- [32:56]
Emmanuel
Well, Joe just eats chickpeas, so he's like, here's a chickpeas for reference.
- [32:59]
Jason Kebler
There you go. That's it. Just give My.
- [33:01]
Emmanuel
Sorry, I didn't mean to interrupt you, Jason, but that was something I thought of because it's like this. It's also confusing.
- [33:06]
Unnamed Speaker
I think you're absolutely correct. And I just. I don't know. I don't know. I mean, I. I would argue that, like, a lot of the translation stuff is then used for generative AI stuff.
- [33:18]
Emmanuel
Yeah, definitely. Some way getting fed into, like, AI data sets.
- [33:22]
Unnamed Speaker
Yeah. And it's like, you can also now, I mean, I would consider 11 labs, for example, which does, you know, like, voices and translation and things like that, to be generative AI. Yeah, but 11 Labs also does translation for that. And it's like there's a mix of just like generative stuff and not generative stuff, like the actual translation of it. I don't know. It's like, I think that they're. I think inquiring minds could disagree on where the line is.
- [33:57]
Emmanuel
I'm sure that our listeners who are way smarter than me on a good day will let us know.
- [34:03]
Jason Kebler
I'm sure, and I hope so. Let's leave that there because we have another conversation we want to do, which is more about what's happened in media and journalism over the past year in regards to our own publication, but obviously many others as well, big and small. We'll be right back after this. All right. And we are back. Back to talk a little bit about journalism and I suppose sort of the business of media. We launched. Shit. How long now? 12, 13, 14 months, 15 months, something like that. There's been a lot. And we came obviously. Well, not obviously for maybe people who don't know. The slow bankruptcy of Vice Media and the technology section of that called Motherboard. And then we leave and then we make this. There's been many other smaller independent media outlets as well, you know, Hellgate and then a massive one defector. Then you have remap, also from Vice. You have Aftermath, who also cover games. I think, Jason, the first thing you have here is just the real collapse of finding readers. What do you mean about. What do you mean by that? And how does that apply to us and others?
- [35:50]
Unnamed Speaker
Yeah, I think that for the last several decades, it's like, let's say you want to find. Let's say you want to get traffic, which was the goal of pretty much every media company that has really ever existed on the Internet for a long time. And it was like you tried to get people to read your stuff via Google by having search engine optimized articles, and you tried to get people to. To read your stuff via Facebook. And then Facebook obviously like kind of died as a traffic refer several years ago, but Twitter kind of, you know, replaced that in, in many ways. And Twitter was always a place where it never drove like tons of traffic necessarily, but it served as it had a very like influential readership and user base that would then kind of like push the articles everywhere else. And the way that you got traffic on the Internet was not like a mystery really. It's like you hit. You won the Google lottery by getting to the top of a Google search result or getting on Google News. You won a social media lottery by doing something that was shared very virally on Twitter or Facebook and then you like collected pennies of ad revenue. Like that was pretty much the model. And that model collapsed a few years ago. The ad model collapsed first where, you know, you were getting fewer and fewer pennies per page view. And then that sort of led to the bankruptcy advice and led to really bad outcomes for like buzzfeed and Huffington Post. And you know, Vox is still around, but like many, many, many new media startups sort of like died or became shells of what they once were. But now, like, really it's not clear how you get a big audience. And increasingly I would argue it doesn't even matter. Like you don't need to get as large of an audience as you once did because what has happened is Twitter has become a disaster. It doesn't drive nearly as much attention as it once did. Unless you are sort of like actively trying to be a like right wing Elon Musk engagement more or less or something. Yeah, yeah. And then it's like Threads is a mess also. Blue sky is exciting and interesting, but it's also very new. And then Google has gotten way less useful because Google first of all is filled with like affiliate marketing links and just like all these little boxes that try to keep you on Google, like the knowledge panels and all of that. But then there's also the artificial intelligence answers at the top which we've talked about before, but. And which went away for a while because it was just giving incorrect information all the time, but which are now back and which I would argue probably are still quite bad, but are there. And like there's not the backlash to them that there was because people can only stay mad for so long. And so it's like a lot harder to build an audience. And I think, I'll be very honest with you, I think we launched at a very good time because when we launched, Twitter had not yet fully collapsed. And I think that we were able to get sort of like a first round of readers just by doing the same, like I'll tweet my story and that's it. But even just in this year, it's like the Internet has become a much more fractured, siloed place. So I guess let's start there. Maybe like we've been able to get an audience, but I think the way that you do that has changed significantly.
- [39:45]
Jason Kebler
Yeah, I'm most worried for. Well, on one side of the spectrum you have sites like ours which are look to be real. We grow steadily and we want to grow sustainably and responsibly. But we're niche. There was a really good interview on Dakota Nilay Patel's podcast and it was about how you have these outlets which are basically for nerds, which I took as a compliment. And I think we were actually mentioned in the same breath or at least a couple of seconds later or before. But you have these very niche publications which are serving to a smaller, more dedicated audience. And then you'll have the ones on the far, far other end of the spectrum like the New York Times, which is basically a tech company now with a newspaper attached where they're just buying wordle and really developing their products all over on that side. I'm worried about the ones in between, Louisiana Times, you know, where they're losing their staff and all of that sort of thing. And then we have a billionaire owner who's saying we're going to put an AI powered bias meter into our articles. Like really, really crazy shit. The buzzfeeds as well, where you have it was one of the best investigations teams in the entire business. What do you have to replace that with when it's either just one or the other? And I think you're right in that. How do you even go about building an audience that can grow to that size? If all of the traditional, or at least the distribution models that you relied on and you understood how they worked when they're all falling around you. I mean our idea that, well, our solution that we go on and on about is that heavily emphasize email, right? That we, we want to bypass all of these platforms in case they do fall apart. And we just have this email list of people who want to give it to us and opt in to receive our articles and that sort of thing. And who knows, you know, maybe years down the line we could grow into a more of a medium sized media outlet. I have no idea if that's possible. I don't even know if we want that right but there hasn't been an example of that happening. I mean, except defector. How big is Defector now, Jason, off the top of your head?
- [42:07]
Unnamed Speaker
It's like 40,000 paid subscribers, but they have probably 25 employees, something like that.
- [42:13]
Jason Kebler
That's a lot. That's a lot.
- [42:15]
Unnamed Speaker
It is a lot. Yeah. Like, I've mentioned this on some of the podcasts that I've been on where they've sort of talked to me. And it's like we've created four journalism jobs, and there have been thousands and thousands of journalism jobs lost, especially at the local news level. And I think that's really scary. That's still not a solved problem. But I do think that the quote unquote market can support many, many more of us. Like Many, many more 404 medias. If you're in a specific niche, the tools to do it have gotten a lot easier. A lot more like off the shelf, accessible. And the economics of it are not so crazy that it's like, can you get a thousand or a couple thousand people to pay for what you're doing? And I think that that is doable for a lot of people, a lot of journalists. But I think the way you do that, it. It has to be, or at least the way that we've found it to work, is you have to be doing, like, original reporting. You have to be getting original scoops that you cannot find elsewhere. Because for a while, you could just, like, aggregate other people's work and hope that your version of the story got read very widely. That's like, not. I don't think that works anymore. Maybe there's some version of it where it still works. If you're like the Daily Mail, because you're just like. You have hundreds of journalists or writers doing it well, and tech websites still.
- [43:45]
Jason Kebler
Do it, you'll have, you know, the Gizmodos and the Mashable and all of that. And sometimes they do do good original reporting as well. And sometimes aggregations actually add a little bit of new insight. But they're still very much doing that model. And like, we did that motherboard sometimes. I don't think to the same extent. Right. But, you know, if a report came out and somebody already covered it, maybe we would quickly hit it, but we just don't do that anymore because why would someone come and give us money if they can just go read it elsewhere? As you say, you need original stuff to actually provide people.
- [44:19]
Unnamed Speaker
You want to be the ones getting aggregated at this point, or you want to be the one who like there's newsletters out there. You want that newsletter writer to include your article in it. And I think that that is what has really worked for us as far as growing our audience is that people are saying, well, you can come to 404 Media and you'll find stories that you cannot find anywhere else. And I think that's, that's a whole different skill set than a lot of places have been training their journalists for the last decade. Like there are many, many places, like I see you have on here the messenger, which is like a disaster of a website that was. They, they leaned really hard into like we're just going to aggregate stuff. We're just going to like see what other people are doing and then we're going to write a 300 word version of it, like over and over and over again. And it failed within a few months last year.
- [45:14]
Jason Kebler
And they hired aggressively dozens upon dozens of journalists from very high paying jobs. I think I got pitched at one point and then it's like, yeah, it just, I mean I'm not going to pretend, oh, I knew that was going to happen. I know I did. I wasn't really thinking about the business there, but they were trying to do a Ray Ray old model in a time where it's like you can't do this anymore.
- [45:36]
Unnamed Speaker
Right.
- [45:36]
And there, there are people who work there. There were some people like when that shut down, said, like, I worked here for eight months, I wrote 300 articles and I'm not proud of a single one of them. Like I didn't do any original reporting and therefore I don't have the skills to do it. You know, anyone can do it if, if they sort of are taught how to do it and get the opportunity to like learn how to do reporting and stuff like that. But a lot of these media outlets that have failed in the middle didn't do a lot of original reporting. I mean many of them did, but a lot of them did not. And I think that is, that's bad. That's like a bad sitch.
- [46:15]
Jason Kebler
Yeah. You've also got here human curation that cuts through the noise. Like what, what, what do you mean by that? Are you referring to like newsletters that can provide value? Because it's almost different. You know, you might have aggregators who will just quickly aggregate a news story and you have newsletters who really keep tabs on so many different sites and other newsletters potentially around the web all at once and then pick out what really matters and provide value that way. Is that what you Mean, like, what do you mean by.
- [46:48]
Unnamed Speaker
Yeah, I mean, I don't know if we're at peak newsletter. There's like tons and tons and tons of newsletters. But I think that as it becomes harder to wade through Twitter or to wade through Blue sky or Google, people are relying on influencers that they trust, like, you know, God forbid, like YouTubers, Twitch streamers, Instagram accounts, and TikTokers that they, they're like, I like what they post, so I'm gonna follow them. And, and I like their perspective, so I'm gonna follow them. Like that's something that we've seen. And I think that we have not done that much on those platforms, like we. We've started doing more recently. But I think part of the reason we have this podcast is, is it's a different medium and a different way of talking about our stories. But what. When it comes to like, curating stuff, I think it could be newsletters where it's like, I can't find shit on Google. I don't have time to doom scroll through Twitter all day. So I'm just going to read this one tech newsletter that has 20 good links in it and I'll click what I want to. There's a. There's a newsletter called 1440 that Sam is a huge fan of. Number one fan that, that you said that. You said that.
- [48:06]
Emmanuel
I like them. I read them.
- [48:08]
Unnamed Speaker
Yeah, that's just like, here's a couple links and there's like not that much information about it. And they have like 4 million subscribers. I haven't read that much 1440, so I don't know if it's good or bad really, but it's just very straightforward.
- [48:24]
Emmanuel
It's just like, it's not. There's not a lot of angle, opinion, type anything happening. It's also, like you said, it's just like short. It's almost like an RSS feed. It's like that short. It's really just like the headline and then like a little tiny blurb about what's going on, which I appreciate because the Internet is busy.
- [48:44]
Unnamed Speaker
Return of RSS is, is. Has happened as well. Sam, you're. You're. You're back on RSS in a big way.
- [48:50]
Emmanuel
Kind of back. It's not a big way. I'm like, I'm. I struggle to find the good. A good RSS reader, honestly. I think they're all a little bit bad or I'm using them wrong. But like, Feedly is. I struggle with Feedly and I. In our reader. I cannot figure out. But yeah, I mean that's, that's what I want though. I want that sort of way of consuming, for lack of a better word, like what's going on on the Internet. Because everything else is just so overwhelming right now.
- [49:21]
Jason Kebler
Why do you want it in rss? Because you can. Is it that it's presented in a much more clean way? You don't have to go out and like search and keep scrolling through Twitter, like, why do you want your news like that?
- [49:36]
Emmanuel
I mean, it's like the, like consuming it on social media has always been really bad. It's like a bad idea always. But like lately it's just like everyone is trying to get everyone else whipped up all the time and I am exhausted by it. So just having the information just straightforwardly put in front of my face and let me figure out how to think about it would be my preferred mode of like consuming the news. Short of like if I really, if I wasn't doing this for a living, I would just be like one of those people who reads the paper. One of those people just get the newsletter. Get the newsletter. The newspaper delivered once a day, read whatever is on it and then put it down. Sounds nice. So trying to get closer to that.
- [50:33]
Jason Kebler
Here is a human curated list of stories that we think are important from our mastheads. It's basically human curation, obviously. Just making an editorial decision on what stories are important. And you're delegating the picking of it to the paper, basically.
- [50:49]
Emmanuel
Yeah, I mean, yeah, there's always an angle, obviously. There's always like even like 1440, I'm sure has they pick and choose what they're going to show me in any given newsletter too. So it's not like it's totally, you know, unbiased or whatever. But yeah, I don't know, it's like it's all just so scrambled up right now. Maybe we should do one. Maybe we should do a newsletter like that.
- [51:11]
Unnamed Speaker
We should, we should. I have an idea for it and I think it's really good. I don't know if the technology exists to do it yet, but I think it's very good idea. I'll tell you later.
- [51:22]
Jason Kebler
Yeah, maybe not right now.
- [51:24]
Emmanuel
A little bit of foreshadowing.
- [51:25]
Jason Kebler
Yes. But you mentioned RSS and I think that just brings to something I just wanted to just mention because I think it's interesting in that the Verge, very long running technology news website that we often have crossed paths with over the years with former colleagues going to work there, et cetera, et cetera. They've put up a paywall now on a ton of their coverage. Also, I'm pretty sure if you pay, you get access to the RSS fees, which is very similar to ours. The vast majority of the content on our website is payable, and if you want to use an RSS feed, you're going to have to get one that bypasses the paywall, obviously, and you get that once you pay. Verge is doing something similar. I just think it's interesting and I wish them obviously all the best of luck doing it. I will be watching very, very closely how successful it will be for them. Obviously, I have no inside knowledge and only going to understand and see what they share publicly, but I look at something like TechCrunch, which is very much an industry leaning publication. They had a subscription product, kind of like an insider sort of thing, and I think they sunsetted that eventually. I think. I was actually at an event where TechCrunch spoke, I think the head of it, and they mentioned that it was good, but maybe not as successful as it could have been or perhaps should have been. So that didn't happen. I wonder if the Verges will keep going. No idea. Again, wish them all the best of luck. I just find it interesting, sort of in the same way they were talking about. On one side you have the tiny publications like ours, then you have the big ones, right? On the other side, those people in the middle. I think it can be really, really hard to launch a paywall or a subscription product when you haven't done for so, so long, which is why we had the benefit of basically starting fresh. And immediately we were like, we need your subscriptions, you know, and this isn't. I'm not saying this is the case with the Verge, but say something like Vice. And if you go to Vice, you can sometimes see at the moment they're asking for subscriptions. There's a difference between giving $10 a month to us, because there's our four faces here on this podcast, and it's literally going to us, right? And then we're the ones who are choosing what to do with it and to hire people and maybe invest in new projects and build the RSS feed, that sort of thing. That is so different to giving 10 bucks to a faceless corporation. And you have no idea where that money is going. And I just don't know how successful a company like Vice, for example, is going to be turning on a subscription, you know, And I definitely think more people are going to. More companies are going to try, though. Crucially not more people, more companies are going to try it.
- [54:38]
Unnamed Speaker
This is the very interesting thing about Waypoint plus and REMAP Radio. So Waypoint was Vice's video game section that had a really dedicated podcast listenership and community and readership. And I managed them along with Emmanuel. Eventually, they started off as their own thing, and there was a long period of time where Vice was like, what are we going to do with this, with these people? Because they have this really dedicated audience, but they don't have a huge mass audience. And we decided to launch Waypoint plus, which was a subscription product, and we launched it, and their. Their community supported them. Like, thousands of people subscribed to them. And that was nominally supposed to, like, protect their jobs, et cetera, but, like, I don't really know how or what to say here, but it's like, it was successful, but that success didn't really, like, come back to Rob and Patrick and Kado in the form of money. It came back to them in the form of, like, they didn't get fired for longer than they probably would have. Like, they. They kept their job, their jobs longer than they would have been able to if Waypoint plus did not exist. But, like, let's say it was a wildly successful thing where they were bringing in, like, millions of dollars. There was no, like, plan to, like, pass that money on to them in some way. And it, I guess it just so happened that, you know, they brought in enough money to, like, sort of keep it, like, more or less sustainable for a little while, but eventually they were able to spin it off into their own thing called REMAP Radio, which you should go listen to and subscribe to and support. But it's like every new subscriber that they're getting, every recurring subscriber that they're getting is going directly to them. And that is, like, so much better for them than the setup that they had at Vice, even though, as their manager, it's like, I went into it trying to make it as, like, equitable for them as possible. But I do think that subscriptions have been a part of the media forever. They've come back. They're going to continue to be. The New York Times has subscribers. Like, obviously, like, all of these publications have subscribers. Subscribers. The question is, just, like, when you have someone who's very good at what they do and who drives a lot of subscribers, it's like, how are they compensated? How does that work? Emmanuel, it looks like you were gonna say something.
- [57:23]
I mean, I don't know what to say about The Verges subscription product either. Other than like, many talented people work there and good luck. It's like, I hope that the people who are creating the value at that company get to keep their jobs and are rewarded for building a subscription business. Is that if that's what they are doing. But yeah, I guess just to repeat what Jason said, I do know what to say about our own subscription, which is every single dollar that you send us, we all see it and we all decide what to do with it. And I think that's why people are supporting us and why they supported like all these other people on Substack. Like, intuitively, I think it makes a lot of sense to readers to be like, okay, there's one person behind this substack or there's four people behind four, four Media. And I like their reporting and I trust what they are doing. So I am putting money in their pocket so they can decide what to do with it. Which is exactly what we're doing. We had a conversation about doing a subscription business for Motherboard and for Vice back when we were there. And I think the reason that it worked for Waypoint is because by their nature and by their design, they very much built their own sub brand that had its own identity and that people really identified with. And that's why people subscribe by. But they did that by overcoming the fact that they were working for a large corporation. And when we were talking about doing it for Motherboard or for doing it for Vice, we were like, there's no way, like, there's no way that we can make some pitch to our readers to be like, hey, we know that. You know that there's like a bunch of executives getting paid millions of dollars above us. Why don't you give them money so we can keep our jobs for another 16 months or something? You know what I mean? It's like, that's not a good pitch, you know, and that's why we never, never even like, attempted to make it. This pitch makes a lot more sense. And thank you again for, for, for subscribing.
- [59:36]
Jason Kebler
Yeah, I guess there's just. The last thing I'll say before we wrap up is like, I was going to mention sort of influencers and that sort of thing. And I don't just mean like, oh, stereotypical Instagram influencers selling whatever niche they have. I mean, more what Emmanuel was getting at, that you have all these people with individual substacks, you have these people with smaller outlets like ourselves. And it's something I've had to sort of come to terms with over the past year and a half or whatever. But elements of our branding do represent influencers in some sort of way. We're very serious journalists who do a lot of really serious, hard, difficult, impactful work. But there is some sort of crossover there. I won't call us influencers, I would still call us journalists, obviously, but we have to have our faces out there, literally, which is very weird for me as a very privacy centered person. People are, as you say, Emmanuel, sort of buying into the brand. And when it's a substack, maybe they're like, I don't know, they have a lot of opinions that you agree with or maybe they provide actual interesting analysis or maybe sometimes they just have a really strong voice. I don't know if this is necessarily true or not, but if I had to say I had a brand and I'm only talking for myself, just personally, it's like if my brand would be like, well, he's just a very middle of the road journalist. That's almost the anti brand of it. Like I don't have a voice. I'm very, very. Oh no, you can't do that. Wagging my finger, trying to be an ethical journalist or whatever to the point where it almost becomes a little bit of a caricature and actually does become a brand really in and of itself. But as you say, people want to support individuals or small groups of people. And I say that both as somebody who is running this business with all of you and as somebody who subscribes to a ton of independent media. Because when I go and send money to somebody, I know it's going to them and it's not going to some faceless corporation, you know. And again, that middle ground of these middle size newsrooms, I don't know whether they have the benefit of that.
- [61:48]
Unnamed Speaker
I'm trying to become an Instagram influencer, sharing our work on my own Instagram page. Not successful so far, but you can find me on Instagram. Please follow.
- [62:01]
Jason Kebler
Yeah, sure. All right, how about we'll leave that there. As I said, this was a pretty different episode. Hope you enjoyed it. We'll be back in the new year, obviously with a ton of original reporting and scoops and all of that.
- [62:17]
Unnamed Speaker
So with that, Happy New Year.
- [62:19]
Jason Kebler
Happy New Year. And I'll play us out. Like I didn't do the intro, but I feel like I should do the outro, so I'm just going to do that real quick. As a reminder, 404 Media is journalist, founded and supported by subscribers. If you wish to subscribe to 404 Media and directly support our work, please go to 404 Media 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. Not this week though. 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 404 Media. We'll see you again next week, slash next year.