Transcript 0:00 [on-hold music] This week's episode of the Rebooting show is brought to you by Beehiiv. I wanna tell you about something I think a lot of you will find useful. 0:16 When you are running a media business, email is often the backbone, and getting it right is, unfortunately, usually harder than it should be. Trust me, I know this. And that's why I've been so impressed with Beehiiv. 0:28 This is a platform that was built by the people who scaled Morning Brew to millions of readers, and they created the platform that they wish they had when they were just starting out as publishers. 0:38 So it's no surprise that companies like Time and the Texas Tribune are already running their newsletters on Beehiiv. It is designed for serious media operators, not marketers bolting on a newsletter as an afterthought. 0:51 And here's the thing, if you're running a publication with more than a hundred thousand subscribers, Beehiiv is built for you. 0:57 Their team has scaled newsletters at places like The Information, Puck, and Morning Brew, and they understand the challenges big publishers face, from seamless migration to enterprise-grade deliverability and monetization. 1:07 Beehiiv makes it easy to grow, retain, and engage audiences at serious scale. If you wanna see what your next stage of growth could look like, go to beehiiv.com/trb. 1:20 That is B-E-E-H-I-I-V.com/trb. And you can meet with Beehiiv's team of growth and newsletter experts today. Again, that is beehiiv.com/trb. Thank you to Beehiiv. Welcome to the Rebooting show. I am Brian Morrissey. 1:38 I am joined this week by Doug Leeds. Doug used to run IAC Publishing for many years. He's a, he's a veteran of digital media. 1:46 That's what, that's what we say, Doug, if someone's got gray hair, we say, "I, I, I'm a veteran," seasoned veteran, that kind of thing. I'll take gray hair. That's- Yeah, exactly... that's actually an improvement. 1:55 But Doug is now the CEO of RSL, okay, Really Simple Licensing, which is an effort to address what might be the definitive shift of our time in digital media, which is the rise of these AI systems that ingest all kinds of content, including publisher content, obviously, at massive scale, and return answers to people, which one of Doug's former companies, Ask, formerly known as Ask Jeeves, promised to do, but AI wasn't quite where it is now, right? 2:27 [chuckles] Like... Still, the idea was completely right on. Yeah. So basically, RSL is a... And Doug's gonna correct me if I get this wrong. 2:35 It is a, a standard that lets publishers set the licensing terms and prices for how AI systems can crawl and use their content. 2:44 It is basically an extension of the logic of robots.txt, where, you know, robots.txt was the little code that you could put on web pages and told search engines, "Don't crawl this stuff." 2:57 I once-- A developer once put it on every single page of ours at a company, at a publishing [chuckles] company I worked with. We couldn't understand why our search traffic had collapsed. [chuckles] That was the reason. 3:05 So it's a very important part of the open web and the history of the open web, which we will get into. But this is all part of a, 3:15 a challenge, obviously, that's out there, and then there's different attempts to try to come up with a new economic model, basically- Yep... for the open web and, and publishing. 3:24 Because the old bargain was, "You can crawl our content, you can take snippets of it and publish it, but you're gonna send traffic back to, to us." Right. And Google was, was the biggest source of that traffic. 3:37 You know, they were gonna monetize it through ads, and Google controlled the ad system, so it was an uneasy bargain. People complained every now and again. 3:45 Everyone likes to complain about the landlord, but, you know, overall- Mm-hmm... I think it worked. And what we have now doesn't work. So, you know, the question is what is gonna replace it? 3:57 Very happy to have you here to, to walk me through it. Did I get anything wrong in the, uh, intro, Doug? No, you, you just missed a piece. [chuckles] Okay. 4:04 So, like, I can supplement it because- This is why I usually do the intros- Yeah... afterwards, but go on. This is good. No. Well, you, you, you got it right. We, we... 4:11 RSL is a standard, an open standard, so that you can say what your terms are for using your content, and that can go into robots.txt. It, it can do more than that, but that's- Yeah... originally what we, we built it for. 4:23 But we also are building on top of that a collective rights organization, which is distinct from the standard, which allow... 4:30 It's one articulation of the standard and allows all the publishers and creators across the entire web to get together- Right... 4:36 and say, "Okay, we'll use one agreement to license our content to AI companies," and it solves a bunch of problems, efficiency and a whole bunch of things, which we can get into. Yeah. I'm happy to tell you about it. 4:47 All right. Let's sequence though. Let's get into- Yes. Let's do it... 'cause this stuff can get, can get wonky, but let's start, like, big picture- Yeah... right? I mean, you've been in this industry for a long time. 4:57 First of all, explain the bargain that underpinned the open web. Yeah, for sure. So, as you said, I ran a search engine, so I was familiar with that. 5:06 Ran media companies, so that there was a bargain between search and media companies, right? Or anybody to say, "You can use my content. You can access my content. 5:14 In the search case, you can crawl my content and index it in exchange for giving me traffic when someone might be interested in that." And Google gives you... Had all the... 5:26 'Cause all they gave you was a list of, of links that were relevant to your search, and you would go there, I'm describing search, everyone knows what search is, and find the content on the page. 5:36 So that's what the, the, the bargain was. You can take s- my content and use it for the purposes of giving me back traffic when you determine that a user is interested in what I have. 5:46 ThatHas gone away or is going away rapidly with a better product, which is AI. I don't have to search on a page for what I'm looking for. 5:55 I don't have to search on the Google website for what's the right content to click on. I don't have to search when I get to the website about what content do I, did I want, was interested in. It just gives me the answer. 6:04 And you're absolutely right, that was the idea of Ask Jeeves, was just give me the answer. And this is actually doing- Smart snippets. What was it? Smart snippets? Smart answers? Smart answers. Smart answers. Wow. See? 6:13 Yeah, yeah, yeah. I go way back, Doug. Wow. Thank you for that. I'm season two. That's... [laughs] Not a lot of people can remember Smart Answers, but that's great. I know. 6:20 Which actually started with a, with, [chuckles] with actually a, a room full of people writing smart answers. 6:26 And ultimately we had technology- There wasn't a lot of AI behind the Smart Answers, is- But there was no AI at that time. Okay. Exactly. 6:31 So that bargain of like, okay, we'll use some of your content, and we'll be an intermediary, and we'll tell people which is the most relevant, but then you get the traffic and you can monetize it, as you said, with ads, but with subscriptions, with commerce- Yeah... 6:43 lots of ways to monetize traffic when humans come to your site, that is going away 'cause you get the, what you need at the AI interface- Sure... 6:51 and you don't end up going to the source of the traffic, and because you're not going there, they can't monetize you. You're right. That's the problem. 6:58 And so right now that i- i- so the situation right now is that's fundamentally broken because obviously publishers are seeing less traffic. Some of that is AI related, some of that is just Google. 7:09 Google has been changing m- more rapidly in the last several years than it had been. It, it became less reliable. I used to, I used to joke with... 7:19 I used to tell Neil Vogel, I'm like, "Ah, you don't wanna be dependent on any algorithm." 7:23 And he's like, "Yeah, but this is the mo- if you're gonna be dependent on one," this is back in the Facebook era, he goes, "You wanna be dependent on this one." And now it's less reliable. 7:30 Some publishers are seeing, you know, really significant, have seen s- really significant declines in traffic. And the question, the question is what replaces it? Now, I've talked with Cloudflare on this podcast- Yeah... 7:43 I don't think I've had Tollbit on. But there's different ideas floating around, and, and these are not... 7:48 Give, give me your overview of the sort of landscape of concepts and companies that are looking to address this issue. Yeah. And I don't think... They're not, like, all opposed. There, there, there's, there's- No. 8:02 They kind of all can- In fact... are, are part of the same ecosystem. That's right. In fact, uh, Cloudflare is a supporter of ours, and we're working closely with them. 8:10 So, and yet- They're a CDN for those- They're a CDN... who did not catch that. And that's a, that is a critical piece of technology- Yeah... 8:16 that basically compresses and allo- uh, basically allo- it handles all the traffic for, for the i- internet. I mean, Cloudflare doesn't handle all the traffic. It's one of a few CDNs, but it is a leading CDN. 8:28 It's a leading CDN, and they're working with us, as is their competitors, Akamai- Fastly... and Fastly. Yeah. Okay. Yeah. Exactly. To... And we'll get into this, I think, a little bit. 8:37 Your question was, what are the models out there? And there are a bunch of models out there. You mentioned Tollbit. 8:43 Th- A- and Cloudflare also has a sort of like, we'll block you, and then we'll charge you to enter the site. And it's like pay per crawl. And that's a common... Pay per crawl. Pay per crawl is a model. A- and it's blunt. 8:54 The way I understand it is it's blunt. Like, I mean- It is blunt... search worked because, you know, the, the click prices were all different, but this is just, you know, it's, it's setting, like, one price. 9:04 Well, it could be setting one price. It doesn't necessarily have to be setting one price. Okay. But you start getting into a lot of difficulty if, in setting prices. 9:12 Look, setting prices is one of the biggest difficulties with, with pay per crawl because imagine yourself as an AI company, and you're like, "I don't know whether- Yeah... 9:22 the content on your site is gonna give me value or not until I crawl it and determine if it's useful to me. So what price am I willing to pay before I know what content I'm gonna get?" Very low, right? At that point. 9:34 I don't know what's there, so I don't know how valuable it is, so I'm only gonna pay you a little bit until I figure it out. On the publisher side, once you crawl, you have my access to my content. 9:43 You can use it a million, billion times. So if you're talking about from a publisher side, I want that price to be very high. You're gonna crawl me, and I don't know what you're gonna use it for, how often. 9:55 I think it's very valuable. If I'm proven right, if I price it where you want me to price it, I'm gonna get almost nothing for it. So there's this, this imbalance in the marketplace- Mm... 10:05 in pricing for pay per crawl, and you called it out exactly. That is the problem with pay per crawl, especially for types of media where you're not crawling it all the time, like constantly, like update, update, update. 10:16 Like New York Times, it's a little bit better. There's other parts of the New York Times. But just for like the, the top news in New York Times, I'm gonna crawl it every day, every day, every day. 10:24 I know the value of that. Okay, pay per crawl might work. But for something like The Atlantic, you crawl it when they have a story like Signalgate, which is like a feature film coming out. 10:35 Lots of effort put into that, huge amounts of cost. 10:38 You crawl it once and you have all the content, and that's the only time you have to pay, it doesn't make a lot of sense, and doesn't pay for the investigative journalism that went into that story. 10:46 So what we're doing is a totally different model under the RSL Collective. 10:52 What we're doing very much, and Stephanie at, at your Cloudflare conversation, had this conversation with you, which was really interesting, about how the world works- Mm-hmm... and, and sort of the Spotify model. 11:05 When Spotify uses content, they pay for it using a collective rights organization called ASCAP or BMI or others. 11:12 They pay one blanket fee, and then the collective management organization distributes that money based on how the content was used and when it was used. 11:21 So basically, 'cause it's music, it's easy to describe, when your song gets played, you as a musician or a rights holder get paid. Yeah. And that's what we're talking about. 11:30 When your content gets used in a result, in a response, then the content owner, rights holder, would get paid. And at that point, it's much easier to determine what the value is of it because it's being used. 11:43 It's delivering value. You don't get paid- Yeah... by Spotify when you upload your song. You get paid by Spotify when your song gets played.Yeah. One thing I would just push on, uh, on that- Please... 11:54 is in that example, I mean, it's a song, right? That, that is being played. It's, uh, that, it's the exact song. And I just know, I mean, we're both seasoned, right? Which means we have gray hairs. Yeah. 12:05 A lot of parts of this industry ha- have, have, uh, run on plausible deniability. All of ad tech, as far as I can tell, runs on plausible deniability. Oh, yeah. And they're... 12:13 And AI has baked into it a plausible deniability in that they're mixing all kinds of different stuff in there, and they're saying, "Oh, well, I mean, we're using the..." Who knows what it is? 12:24 It, it's not really playing a song, it's more like it's listening to a song, and then it's, like, humming a song based on a song it... All the songs it listened to. And, and then they say, "Well, that's impossible." 12:35 And then you sort of throw their hands up in the air and walk away with triumphs. Well, there's a couple things that I'd love to dig into that you just mentioned- Yeah... that I think- Okay... 12:42 are interesting to talk about in this context. All right? One is they're mixing it all together. Let's just talk about that for a second. Okay. They're mixing it all together. 12:51 Absolutely they're doing that, and they're doing that a lot of times because they don't have the rights to not do that. They don't have the rights to the content. They haven't licensed the content. They've crawled it. 13:01 So they believe that if they mix it all together, they have a fair use argument to not have to pay for it. Mm-hmm. I'll put aside the legal argument for a second. We'll come back to it, I'm sure. 13:10 But right now, the idea is they mix it all together because they, because they're not paying for it. So what happens when they mix it all together? One is they spend an enormous amount of money to do that. It's... 13:23 If you wanna just, like, look up what, uh, what Sam Altman said, that OpenAI pays just to respond to the words please and thank you, it's astounding. Okay. 13:32 When you talk about what it takes to actually mix up all this content, it's in the billions of dollars. So they're spending a lot of money to mix up all the content. Why are they doing that instead of... 13:42 Oh, by the way, John Roberts, who works for Neil Vogel at Pete Link, who's- Yeah... 13:46 brilliant on this stuff, had a great analogy I heard recently, is like, when, if you get a recipe that's all mixed up, what you get out of that is food poisoning. And, and that really is the story of mixing it all up. 13:56 Well, he's got, he's got a lot of recipe web pages to protect. Well, he does have a lot of recipe content. So, but yes, I understand John's- But the point is- Dr. John's, Dr. John's point, yes. Dr. John Roberts, yes. 14:04 He's, he's amazing. But the, the... I think the point is well taken, which is- Yeah... that's where you get hallucinations. That's where you get misinformation. 14:12 That's where you get the problems with AI, is because they're mixing it all up. Now, if you ask them, like courts have done, "Why aren't you licensing this content so you don't have that problem?" 14:23 They say, literally in the pleadings in the d- court documents, "We are not doing it because it's not possible to license from hundreds of thousands of different places." That is essentially the problem- Yeah... 14:34 that we are solving because they can now license from... We are now in a position, four months after launch, Brian, we are now positioned to represent over half of the content that AI uses on the internet to train. Okay. 14:46 Over half of the content, they can do a deal with us and get a, a licensed access to that. That's- Yeah... game changing for them. Yeah. 14:53 And, and I think that's an important point because, you know, the music analogy, music is structured very differently from the open web. I mean- Yes... 15:02 music is, and tell me if I'm wrong here, it's an oligopoly, basically. I mean, Spotify, uh, can go to, like, through, like, ASCAP or BMI. 15:12 There's j- there's a few record labels that, that control much of the, the, the industry. They can, but they go through the collective rights because they get everything that way. 15:21 This is where it's a little bit difficult, is that music copyrights are, have performance rights, they have mechanical reproduction rights. They have a whole bunch of different rights- Right... 15:29 involved in musical, and they need a bunch of these rights, so they can get some of them from the labels, but they get a bunch of them from, a bunch of the necessary rights from ASCAP in a collective way. Mm-hmm. 15:38 But they do that because we had Napster, and- Oh, yeah... when Napster came out, right, it was the... It's what we're experiencing now. It's, "Oh my gosh, this is a better product. I don't have to buy a CD." 15:50 Yeah, it was pretty cool. I remember. [laughs] Right? I mean, I was listening to a CD. 15:54 I spent $15 to get one or two songs that I liked with five or six songs I didn't like, and then if I wanted to play another band's song, I'd have to pull the CD out, put the CD back in, play the next song. 16:05 That was terrible. I could make a mixtape, but that took time. This was like, oh my gosh, this is a much better system, a playlist, amazing. All the songs digitally streaming, fantastic. 16:15 The problem was Napster wasn't paying for any of that. 16:18 They took it all, and they gave it away for free, and the music industry was like, "Oh my God, this is going to crush us," because there, it removes the incentives to create music. Yeah. 16:29 Just like the incentives to do anything. And- Do you know what's interesting on that? Just, just to stay- Yeah... on that for a second, is like, at least I remember that, is that everyone kinda knew it was theft, right? 16:40 Yes. Whereas I don't... And maybe it's just because of the way that the internet developed since then, and just maybe society or whatever. I don't think that that is common now. Like, it's, it, it... 16:50 I don't think people see going to an AI, a Ch- ChatGPT or Gemini or anything as being like, oh, maybe in this industry, but I, I- But do you know why, Brian? 17:01 It's because music and the entertainment business was founded on licensing way before technology. It was- And Hillary Rosen was out there saying, "You're stealing this," like, nonstop. 17:11 That helped a fair- And Metallica and others- Yeah... they were saying it. But, but music and entertainment, film and others, was a, is a licensing business, and it's been around 100 years. That's true. 17:21 And it exists as a licensing business. You don't go into a place to listen, like you do with Spotify now. You go to Spotify to listen to music. Yeah. That wasn't really a thing back then. You went and bought music. 17:32 You went to a store and paid for it. It was all licensing and no technology, and technology came and disrupted it and said, "Oh, wait, we have a way to distribute this now without having to purchase it all at once." 17:45 You can buy a subscription to Apple Music. You can buy a subscription to iTunes at the time. You can buy a subscription to Spotify. You can get all of it at the same time. 17:52 That was new, and because that was new, it was breaking this idea of, oh, I pay for music. The webHas been open, as we talked about. Mm-hmm. And you can have anybody walk in. 18:04 There was no licensing infrastructure for the web. It was an open model. We call it the open web, because you can go anywhere and, and consume it. The difference was that was based on humans going there and consuming it. 18:16 Humans being there, consuming it themselves, led us to give them ads or subscriptions or commerce, ways we would monetize it. 18:23 Without the humans getting to the content themselves, there's no monetization, and it is in fact stealing. 18:29 It is saying, "Oh, the way you get paid for this," just like you got paid for buying a CD, "the way you get paid for this is by having humans come, and we're taking that away." 18:36 We are basically setting up a, our own store outside of your store, selling your stuff, and not paying you for that, and that is stealing. We, we can see it. Courts are starting to look at it. 18:50 There, there's two issues on the legal front. I won't go too far into this. It's can they take the content and train on it, and then can they use that content? 18:59 And the training on it has been this fight about whether that's fair use. Mm-hmm. But the use of it, especially where that use undermines the value of the original content, that's infringement. 19:10 And wait, explain the, explain the difference between training and use a little bit more. Sure. So I know, I understand, like, the, the stuff that went out the door, and I mean, they're basically these AI... 19:19 L- it's reading, you know, that's what Laya said. It's like a machine, a robot is reading a lot of content out there, every piece of content. Yeah. That stuff is, is gone. 19:28 It's already read everything, and it's continuing to read. But what do you mean use? It's diff- it's different. It's slightly different points, and one's a legal thing and one's sort of- Okay... a practical thing. 19:37 I apologize to your listeners, because I used to be a lawyer, and so- Uh-oh... sometimes I drift into... Yes, exactly. So, so just, you know- I'll pull you back if you get too far. Pull me back, please. 19:45 Please pull me back. But the idea behind what the arguments are about why this is fair use is that they can take the content, and just like if you went to a museum and you saw a bunch of Picasso paintings- Mm... 19:58 and then you went home and painted something inspired by that, that would be fair use, right? Oh, I, I learned about Picasso. I painted something on my own. That's a fair use of what you saw. That was visual. I learned. 20:10 I trained myself on it. I painted. That's fine. If you took pictures of Picasso and you distributed them and got, so, and pa- and sold those, that would undermine the value of the original work. 20:24 That would be infringement. So one thing you can do is train and learn and create new things. What you can't do is use that to undermine the value of the original work. 20:34 And so while the arguments have happened in court that, hey, training is just like looking at a Picasso and painting my own picture, when you do that and you create a picture that is like Picasso and you're saying, "This is as good as Picasso. 20:47 You don't need Picasso anymore 'cause I got this," and you, and then you don't buy Picasso anymore, that's infringement. Does that help? Yeah, no. Or is it- That, that, that- Okay... that helps quite a bit. 20:58 So how does- Okay. So explain then RSL and how it, it will address this. 'Cause RSL, as I said, it's, it's an outgrowth of RSS, which was a- Yes... foundational technology. 21:10 Some, there's a lot of RSS truthers out there who are like, "You know what? The web went to shit when we sort of went away from, from RSS. Google Reader died for our sins," et cetera. But explain- Yeah... 21:21 the role of, of RSS. Well, look, RSS is a great foundation be, for the web, but also for us, because my co-founder, Ekkart Walter, was a co-creator of RSS. Yeah. 21:31 He, he created RSS back in the '90s when he was at Netscape and brought it to Yahoo, and he came to Yahoo, and a lot of that i- for, for good or bad, but really for good, and for podcasts still for good, 'cause most- Yeah, no... 21:42 podcasts are distributed. I still put, you know- Yeah... you, you have a podcast, y- you send the feed to lots of places. That's right. Exactly. And by the way, that's the model. You're still using the model, right? 21:54 The model is you send your feed out and they listen to it, and then you get paid from that listening, 'cause humans are listening to it. 22:01 And so you can have a sponsor, and they can pay for an ad, and they reach the listeners, and 'cause humans are consuming it. 22:07 But imagine, which is happening, that you distributed, uh, your podcast, it summarized everything, and it said, "Here's what the podcast is. 22:15 In fact, we'll reproduce enough of it so you don't need to listen to the podcast, and you can pay us for that." That's just stealing. So how do we, how do we change, how do we change that, right? 22:27 The way we change that... By the way, the reason why Google and Ya... 22:31 I mean, and, and, and Meta and OpenAI and all these companies are saying they're doing it this way is because, and they've literally said this, there's no other way to use this content. 22:42 There's no way to license h- the open web was created and millions, billions of webpages are out there. Right. How do we license everything? It's impossible. That's the problem we're solving. 22:53 We are saying it is now possible to license everything with one deal. Mm-hmm. Because we will do one deal, one blanket license deal, and it can be for all the content on the web. 23:05 And then if you, or if a mom blogger, or if a person who writes their own recipe wants to participate in this, they sign up with us, and then when their content is used, they get paid. 23:16 They get paid because, let's say, OpenAI pays us the licensing fee. We look at when their content was used, and we pay them based on that. Right. Does that, does that make sense? Yeah, 100%. 23:26 And did I answer your question? I'm, I may not have answered your question. No, no, no, you answered my question. All right. I think w- my next one, and we had discussed this earlier. Yes. 23:32 This is where [chuckles] we're going is, okay, great. Like, you're kinda calling bullshit, honestly, on them, because, like, they're saying, "Oh, we'd love to pay people, but, like, we don't-" Yeah "... 23:41 we can't actually do it. Oh, that's too bad." Yeah. Like, "We can send, like, you know, rockets to Mars and all this, but we couldn't possibly pay for this content that we're using." Yeah. 23:50 "That would be too hard, technically. How would we do that?" Right. Right. Maybe I'm a little bit cynical about that. No, I think that's a fair point. I think it's, I think it's a well-earned ear- earned skepticism. 23:59 So let's just assume this is not a technical problem, right? Yeah. Then you run intoWhere's the leverage? You know, like Dr. 24:07 John's bo- uh, boss, I had a conversation with him, I think it was in, in Cannes, we did a podcast, and he's like, "You have to establish leverage to get them to move," and right now we don't have leverage, and we have to establish that leverage because it isn't necessarily a technical problem. 24:25 They seem like they have plenty of money. We can discuss how much money would go to, to, to publishers. I'm sure there'll be a difference of opinion on that, but it's more than zero. 24:34 It comes down to where is the leverage? What, why, why are they not beating down your doors to, to license the open web's content? 24:41 Well, I, I, I'm not in a position to tell you what those doors look like and how open they are at this moment. Okay. But let me answer the rest of the questions. 24:49 So yes, you can call BS on their arguments, and we've talked about a couple. One is we can't license it efficiently, and that's a problem. 24:58 We've talked about another one, which is that, that they mung all the stuff up for purposes various- Right... but it gives them bad quality, and it costs a lot of money. 25:05 So you, you would think that they have incentives, but your point- But then by the way, they're, they're using these publishers' brands as citations in their results to, to give, uh, to give trust and to give legitimacy to their product. 25:19 I mean, it's, it is a little... When that started out, I was like, "This is a little galling." Yeah. It's really galling. 25:24 Okay, so let's talk about the, that's, the, the incentive system for them to do this from their business model is there. But now let's talk about the enforcement mechanisms and why you can force them to do this. 25:37 One is technical, and we talked about that at the beginning. Cloudflare, Akamai, Fastly, the CDNs. 25:43 Right now, the, the model that Cloudflare has used has been, "We're gonna block you by default, and you're not gonna, we're not gonna let you in." Mm-hmm. And that actually 25:53 has been working in some ways, but it's very hard if you depend on things like search. So, like, OpenAI is blocked way more than Google right now, and the reason for that is because 26:05 OpenAI doesn't give you search, doesn't have search rate. You understand. Yeah. So- No, no, no, but other people don't. The other people don't. Okay. So, so- [laughs] I'm sorry. I'm just like, I- No, no... 26:14 forget that we're on video sometimes. Go on. [laughs] So, so, and by the way, the EU is investigating Google for this very thing right now, right? Yeah. 26:21 So Google has said, "If you don't wanna be in our AI abstracts, that thing at the top of our search results, where we basically summarize the rest and tell you- Yeah... 26:32 that you don't need to go to the search results, if you don't wanna be in that, you can't be in the search results." 26:37 So- So instead of separating the two different, the, instead of having two different crawlers, they have one crawler. Yes. And it's like, you want the search traffic? Well, guess what? 26:46 We're gonna take your content, and we're gonna use it- Right... for the AI overviews. Exactly. 26:50 And the EU is investigating them, and I'm under very good authority knowledge to know that they're not the only ones investigating them for this. Yeah. 27:00 And because they're, they have monopoly power- It seems, it seems like an issue the EU would, would, would feel strongly about. Yes. 27:04 They have monopoly power, or market power anyway, in search, and they're using that to get into AI. But to your point, okay, that means it's very hard to, to block them without giving them a way to come in. 27:19 So that's the first thing. There's a technical blocking, that's great. One more legal thing for you. Okay. Just so you hear this, very quickly. Ziff Davis, one of our partners, sued OpenAI. 27:31 In December, the court ruled on their, their lawsuit and upheld a bunch of their... This is at a preliminary stage, so just can the, can the suit go forward or not? And the court said, "For the most part, yes." 27:42 But one of the ways they said no was under something called the DMCA, the Digital Millennium Copyright Act. Mm-hmm. 27:48 And that, in order to pursue the claim under the DMCA, they had to show that there was technically enforced, that they were creating basically a lock, and that lock was broken. 28:01 And the court said, "There is no lock that's broken. Just putting yourself, saying that you can't crawl me is not enough. You actually have to take technical enforcement measures to do that." 28:11 So the reason why that's, it's great for Ziff Davis to have the, all the claims go forward, and they don't need that one. But the nice thing about DMCA is if you take advantage of that, you don't have to prove fair use. 28:24 It doesn't matter if there's fair use. That's not a c- defense. If you broke someone's lock, it doesn't matter what you do with the information once you took it. You broke in, you broke in, you're liable. 28:33 So all the little guys in the world could use DMCA against the search engines or the AI companies very effectively if there was a technical means to br- to stop them. 28:44 CDNs provide that lock, but they don't provide a key to open it. So it's very hard to say, "Okay, I'm gonna stop you from coming here," it, without saying, "Unless you do something." And that's what RSL provides. 28:56 They say, "Okay, here are the terms for which you can unlock this gate." 29:01 And when you have that, then you can put the lock on your gate, and if someone breaks it or circumvents it, you have access to a legal enforcement regime that's much cheaper than currently only the big players have access to. 29:14 Only The New York Times and Ziff Davis and others that are going after this. Right. Right? So- Yeah, and that's the other thing is, like, how do you solve this at s- at scale- At scale... of the open web? 29:23 Because- Right... the New York Times has leverage. They have, you know, for sure. You know, Ziff might have some leverage. You know, The Atlantic has leverage. 29:31 But, like, regular publishers or just individuals even do not have any leverage. So the DMCA- That's why, by the way, these giant tech companies love creator economy stuff. 29:42 They would really love the idea of, like, oh, how are we gonna negotiate with 10 mill- tens of millions of people versus, like, you know, a small number of big publishers. 29:50 Well, and, and, and this is actually exactly why collective management organizations were created in the first place. Yeah. This is why. 29:58 This was back 110 years ago, radio mu- recorded music was invented, and radio said-Radio was invented too, and so we can play that recorded music, but we don't have the rights to do it. 30:10 How do we negotiate with all these people? And so collective management organizations said, "We will give you an efficient way of doing that, efficient way of getting all this together. And if you don't do it, 30:21 then you have to go license for all these other people." So we're right now in the legal stage of proving that you have to go license this with all these other people. 30:29 But once it's proven, which it will be, then how do you do it is a big problem. And you can't do it with everybody without something like- Yeah... 30:37 a collective management organization, which is exactly what RSL and only RSL is doing right now. 30:43 And we're a nonprofit, which is a very important thing for the media industry because you get companies like a Tolbit or others knocking on the door saying, "Hey, we're happy to help you do this, and we're only gonna take 30, 40, 50% of your revenue to do that." 30:58 And, you know, media companies are... Look, if I was running a media company like I did once, I'd be saying yes to everything and see what happens. Right. But, right, there's no, there's no harm in that. 31:07 But the idea of saying, "Wait, all this value that I'm creating, you're gonna give 30 or 40 or 50% away to someone who is just blocking somebody from getting to that?" That seems hard- Yeah... to believe. 31:20 I mean, publishers are already sort of, y- [chuckles] you know, I mean, they, they felt like intermediaries have been reaching into their pockets in- Exactly... 31:27 in various parts of their businesses, like ad tech, et cetera, for a long time too. 31:32 And because me and my partner, founding partner, have come from that space, we specifically sort, started this and are running this as a nonprofit so that we don't have that issue for the big publishing companies and also for the entire open web. 31:46 Honestly, whoever you are creating content out there, you should get paid, and you shouldn't toss you 30% or 40% or 50% of what your value is. It should cover our costs and the rest go to you. 31:58 And that's how ASCAP works, by the way. Yeah. So just to circle back to, to the incentives, like, so why are they going to come to the bargaining table? I mean, obviously this is, these are global companies and- Yeah... 32:11 you know, places like Europe- Free markets... are, are very different than the United States. Yeah. I don't see just politically in the United States that, you know, they're... 32:19 I mean, obviously the, the Trump administration is all in on AI, China, China, China will be brought out. 32:27 It's like if we, if, if we can't compete with China, and it'll get all mixed up into, into that, what is-- And I don't, I don't know if the courts will, will force this, but do you think that the collection of these kind of pressures will lead to some kind of grand bargain? 32:44 Yes. I absolutely- Okay... do. So look, there's the, there's... We talked about technical enforcement. We talked a little about legal enforcement. Mm-hmm. You're hinting at regulatory enforcement. 32:54 Even if it comes first in the EU, like why does Apple not have a lightning plug anymore? That's 'cause of the EU said- That's true... we all have to have the same thing, right? 33:02 So they complied with that everywhere because it was efficient to comply with that everywhere as opposed to having two different devices. Right. 33:08 So it's okay that it's starting in the EU because it will create the leverage, the, the regulatory leverage to require some, some change. 33:18 But also in your Trump analogy or your, your, not analogy, but your Trump reference- Mm-hmm... of China being the bad guys, it's so easy to show who the bad guys are if they're the ones who aren't licensing content. 33:30 Right. And that's actually the difference right now, and can be the difference in the fu- in the future. 33:36 Good AI companies, American-based AI companies who value human-created content can pay for it efficiently, and bad AI companies, Chinese stealers won't, and by the way, that's the best model for saying, "Those guys are bad. 33:52 As a country, we are blocking them. These guys are good. As a country or as a, as a s- a federation of countries, these guys are allowed." That crosses political aisles. Like- Yeah... not, not, and that's pro-AI. 34:05 And by the way, I'm extremely pro-AI. I'm extremely pro-AI. I love to use it. I love the future of, of using it. I have become so much more productive. We're doing this as a nonprofit, which means we're pretty lean. 34:20 Yeah. I don't think we could do a lot of what we're doing without the help of AI, right? 34:24 So it's a much, like we talked about Netflix, it's a much better product, but it has to be a sustainable ecosystem if we want it to continue, and that's my biggest point about why does AI kinda do this? 34:39 'Cause they understand that. They totally understand that. If they don't pay for the content, ultimately the content will go away, and so will their value of their product. 34:50 So what does a fair, what does a fair system look like, right? Like, I mean- Yeah... I think about like sports leagues and, you know, they're basically, you know, when they negotiate- It's a great analogy... 34:58 with the union, they're, they're saying, "Okay, well, are the, are the players gonna get 55%? Are the owners gonna get 45%? Is it gonna be 55% of the owners, 45% the players?" Right. You know, ultimately, 35:09 there needs to be, you know, the economics in place and how do you sort that? 35:14 Because the obvious thing is, you know, publishers will set a price and, you know, AI companies will be like, "Yeah, we're thinking maybe like a third or a fifth of what y-you're saying it's, it's worth." Yeah. 35:27 So with collective rights organizations, like you mentioned, like a, a labor union is, is not that different 'cause you're representing everybody, right? Mm-hmm. In the labor union. 35:35 You have to come up, you have to bargain, and you have to come up with a reasonable reason why what you're proposing makes sense. Right. There are three pillars of AI, right? 35:45 There's power, there's chips, and there's information, and they're not paying for information right now. So we know that what-- they're paying billions, tens of billions, trillions for the rest of the stuff. 35:56 So we know it's more than what they're making now. We know it's probably less than what publishers on their own would ask for. 36:04 So how do we figure out, your question, how do we figure out what it is?And I'll tell you how we're doing it. 36:10 We've hired economists, and we're looking at the market specifically with the intention of being what's called FRAND. FRAND is an acronym meaning- Huh... fair, reasonable, and non-discriminatory. And that- Okay... 36:23 is what we're looking to be. And we're hiring economists to say, "Okay, how does the market work? What is the value of it? Look, take, take it apart." 36:32 Again, my friend, my good friend John Roberts, has done some of the analysis on token value, and it's actually very... 36:38 The, the, the amount of money that pays for all the content in the world is not unlike Napster, where they were giving stuff away for free. This is already being collected. 36:48 This money, very close to the amount of money that it would take to pay for all this, is already being collected. So- Mm-hmm... 36:54 one reason I'm not saying let's just take the money that's being collected is because I think they're underpricing AI right now. I'll give you my own personal- Why, why, explain that. Why, why, why is that? 37:03 Because, because we're early. And I, I mean, I... We talk about gray beards. Mm-hmm. We've been around, we've seen plenty of this stuff. You price- Yeah... 37:10 low at the beginning so that you capture market share, and then you- Yeah... raise your prices when you have it. So that's the mode we're in right now. I pay $20 a month for ChatGPT. 37:19 They raised it to $100 a month, I'd pay it in a second. There's no question I'd pay it. Would they lose five, four of their, their other customers if they raised it? No, I don't think so. 37:27 But they're basically doing this because they wanna capture market share. It's not based on the value that they're delivering. It's based on a long-term thought of what can I get later if I get a bunch of customers now. 37:40 Which is fine, but that doesn't affect how much they pay for power or how much they pay for processing, and it shouldn't affect how much they pay for information. So how do we figure that out? 37:52 We hire economists, and that's what we've done, and they're working on it right now to tell us what is a fair amount so that we can say to the world transparently, and to AI companies, this is a fair amount. Yeah. 38:04 I mean, Spotify pays, I believe, about 70% out to artists. I think it's higher than that. Six- So, oh, okay... seven... I saw 75% or 76%. Okay. Of course, I think I saw that through AI and through the negotiation. 38:13 I mean, that's, that's quite a bit, so I'm not, like, you know, sure- But that's not-... if the publisher is- But, but, but look, like I said, I'm pro-AI, so here's the pro-AI, AI company, LLM thing. Yeah. 38:22 Spotify is like a mechanical jukebox. Yeah. It's not fair. But, like, all you do is you go there- They're not paying for compute, like, like you say... and you consume it. They're not paying for... 38:29 They're not adding a ton of value. They're adding some value. They're not adding- Right... a ton of value that AI companies add and pay for. So that 75% is not realistic- Right... given just the product, right? 38:41 But- But I, I think understanding who adds what value is... Look, I mean, I always- Exactly... 38:46 any time I talk with, you know, people in, on the programmatic side, the differences between the, what they believe the value is that they are, they are... 38:55 when they enrich a, an impression with some kinda unique data set, maybe it's unique, maybe it's not, but they're claiming that, ah, that's transformative, and, you know, that is where the real value is, and, you know, the place where the ad ran is, you know, kind of, you know, not that valuable. 39:12 There, there will be undoubtedly disagreements about this. Mm. But if we go back to the music model, ASCAP, BMI, they don't represent everybody. They represent almost everybody. 39:24 I believe Taylor Swift has her own deal, and that makes sense, and there can be your own deals in these situations. But everybody isn't Taylor Swift. 39:34 Everybody can't get, "Hey, I need my own deal," type of negotiation with them. They just won't do the deal. 39:41 So we have to have a fair and reasonable economic s- basis for charging for all the content, even if it's not all the content. It's all the content, but the Taylor Swift version of the content. Yeah. 39:55 So when you think back, could this, like, end up m- making the web less ad reliant? Oh, yeah. We, we're going that way regardless. 'Cause you've been in ad businesses, and there's pluses- I have... 40:06 and minuses to being in ad-driven models, and there are a lot of minuses. There's a lot of trade-offs, and we all see it when we experience the web, and, and the open web- It's hard to think of-... 40:16 in particular has suffered from that... anybody who's gone into an ad business where a subscription business would've given them the same revenue portfolio. 40:22 Usually, the ad business is taken up because there's a trade-off. I give you my content for free. You look at an ad. Yeah. Doesn't cost you any money to do that, so I make more money, you make more money, right? 40:33 All that makes sense. 40:35 The question is, it's not that ads are gonna go away, and we've seen this already with Perplexity, which is talking about ads, or ProRata, which is talking about ads, or OpenAI, which is talking about ads. 40:45 It's just where do those ads live? They live where the humors, humans, I don't know what a humor is, where the humans- Mm... interact with content. 40:53 And when the humans stop interacting with content, or do so less on the publisher side and do so more on the AI side, that's where the ads are gonna go. Just like there's ads in versions of Spotify. 41:05 There are ads in versions of Spotify. That's fine. But where does, how do the musicians get paid for those ads? They get it through a licensing deal as opposed to hosting their own advertising. Right. 41:15 But I think it seems like, like publishers, and I've talked with John about this, like what, like- Yeah... they'll end up having a larger licensing. I mean, you see it right even in PeopleLink's of results now. 41:26 Look at Redit. Their, their licensing line is, is, is much- Yeah... different than it was a year ago. Yes. And, and I wonder- That's right... 41:32 how that changes when, when publishers start to operate like almost like wholesale businesses and retail businesses, businesses. It's not like they don't have licensing and there's- Right... different forms of it. 41:42 But look, look, look at the entertainment business we just talked about. Like music is all licensing. You still know who Taylor Swift is even though she makes her money from licensing. That's how she makes her money. 41:51 Well, she ma- she makes her money from, from her touring. I think that's what a lot of- Well, definitely... publishers are gonna be doing. You know? [laughs] You're absolutely right. I know this from the B2B world. 41:58 Like- You're absolutely right. You're absolutely right... I, I always compared, like, being a publisher, it's like-You know, being a musician, it's like your recorded music, you know, is not... 42:08 It used to be, oh, you could go into the, the studio, cut an album, and then, like make a ton of money. And now you look at the, the Oasis tours, you know, they gotta be out there touring and, you know, the... 42:18 It's not like the albums- Yeah... do a job, it's just a different job, and I think publishing- Yeah... is, is kind of similar. And there is a, there, there are some bad parts of the analogy, and this is one of those. 42:26 Yeah. Because, because the rights have been so fractured in music, and because, as you said, there's an oligopoly in the labels who take a lot of the value, a lot of it doesn't go downstream to the musicians themselves. 42:38 We don't have that structure in publishing actually. 42:42 So there's much more likelihood that licensing value goes from the place where the value is created, like an OpenAI, to the place where the value is produced, like People Inc. That- Right... 42:56 there, there's gonna be less intermediaries in there in our model, in, in our, in our industry than there is in music. What would be the... 43:03 W- If you had to make it, obviously you're, you're, uh, you're, you're giving the bull case. Put yourself the other side. Like i-g- what would the bear case be for a, a fair, in quotes, like workable solution coming? 43:18 Well, I think it depends on who you are and how, how valuable your content is. And let's just look at past evolutions. Mm-hmm. There... And w- I think we just saw that the Pi- the Pittsburgh Post- Gazette? Gazette. 43:31 Thank you. Yeah. Pittsburgh Post-Gazette just folded. Would that have happened 40 years ago? Heck no. Right? Because you needed local news, and the way you got local news was by buying a newspaper. 43:43 You had to buy it locally. And now with distribution being, oh, I can get most of that news online, the very hyper local has a different use case. 43:54 It used to be, like for newspapers, I've talked with a bunch of newspapers about this, a bunch who are supporting us, like USA Today, it used to be that the ad, the ad model for newspapers 44:04 was, is there a big mall in town? 'Cause they're gonna supply all the ads. What are classifieds? What happens with them? Okay, classifieds are a big ad source. And all of that changes. 44:17 And so those who become over-reliant on a monetary system that is going to evolve are not going to make it. And the bear case I would say is not for the industry, but for some actors in the industry that don't evolve. 44:33 Right. And I would think, like, you look, everyone follows incentives, right? Right. 44:37 And the incentives of the ad-driven model, particularly when platforms had a choke hold on distribution, was to chase algorithms with s- some fairly undifferentiated content, right? 44:51 If you're a g- a Google, quote, unquote, "Google Discover publisher," look, a lot of your stuff is not that unique really. 44:58 It's l- it's, it's, it's trying to get in front of the line and Google sends a bunch of traffic to you. It might bounce real quick- Yeah... but like, you know, you'll put a lot of ads in front of them. 45:07 I think when you get into a licensing context, I would think the, the incentives shift. Yes. I mean, I think like in the best case scenario- Absolutely... 45:16 the incentives shift to create unique, valuable, essential content, whereas the last- Absolutely... 45:21 system, the way a lot of people operated within it, not all but a lot, was really focused on just trying to get the piece of content to get picked up by the algorithm in some way. Well, the algorithm was a... 45:35 What the algorithm was, was a way to try and determine what content was best. Yeah. That's what the algorithm was, right? 45:43 And then you, once you say, "Here are the rules for determining what content is best, and I only want the best content," then everyone who makes content says, "Okay, I'm gonna understand what those rules are, and I'm going to optimize for those rules." 45:55 So I bought About.com, and that's when I hired Neal Vogel to be the CEO of the business that's now People Inc, right? To run that business. 46:04 That business was based on a version of PageRank that soon went away, which is why Neal evolved the company. 46:10 But that was, hey, if you had one giant storefront of all the content you could possibly want, Google thinks that's more valuable. PageRank is higher. 46:19 So About.com was, was formed under the idea of we'll be about everything, and therefore all the content lives here, and PageRank gives us, says we're, or their algorithm says we're more valuable because of that, so that's how we're gonna operate. 46:34 When they changed their algorithm and said, "You know what? That's stupid," and I agree, all the... The best content doesn't live in one place. The best content lives in different places. Mm-hmm. 46:43 Like, then a- and Neal and I was there and running it at the time, but it was Neal and his team's ideas, let's break this apart. We have really good content. Mm-hmm. 46:52 Instead of putting it all in one place, let's put it in very well. I bought the Investopedia. Let's put it in The Balance. Let's put it in Spruce. 46:59 Let's put it in these brands, because the algorithm then says that's the better place. When we got to social media, it started to be shares. Shares were the way to see if your content was valuable. 47:12 The more shares you had, the more the algorithm said your content is high quality. So what happened? Clickbait. 47:20 Okay, if I wanna share or if I'm gonna get you in there, I'm gonna say something provocative to get you in there to share it out. Sure. Right? 47:27 So this is time and time over again, all the way back from town criers reading headlines out. When you say what is the algorithm for what is valuable content, content producers will produce to that algorithm. 47:42 That's not necessarily a bad thing. The thing is, this is where technology disruption comes from, that someone says, "I have a better idea for an algorithm. This isn't as good as it could be if we did something else." 47:52 And if it's right, then the world adapts. Right. And that is what we're seeing. But it could in some ways, like reward a different type of publisher disproportionally- Oh, yeah... 48:01 to than the last-I mean, 'cause like you mentioned, like, you know, the, the brands, like, breaking apart About into the brands. 48:07 I mean, a lot of those, a lot of the brands, I mean, I had said this to Neil, like, you know, at the time, they were kinda like minimally plausible. Like, I mean, they were plausible within a search context. 48:17 It was like, oh, Verywell. But Verywell didn't... I mean, that's, I assume why, why, you know, Meredith was, was purchased because those brands had more resonance beyond a nice logo, but no history or anything. 48:30 Nobody had, like, independently known of them necessarily. Not nobody, but, like- No. The business did so much better when we went from About to Verywell. Yeah. So much better. And better when you have better brands. 48:41 So we-- all that is right. All that is describing the same thing, which is an evolution of what is the algorithm valuing, and then content producers going after whatever the algorithm. 48:54 I, I'm saying it like it's some- I know... bestowed upon from high. 48:58 Like, but what it is, is, is, is technology people or others saying, "I have an idea," or, "I have a process that can make you get better content than the old idea," and it's embodied in an algorithm. 49:10 And if people agree with it, because it is actually better, then content producers produce to it. And there will be winners and losers along the way. There's no question. There'll be winners and losers along the way. 49:20 So RSL will include individuals too? Hundred percent. Okay. This is about- That's important... the entire web and making sure that, again, like Ask.com, like, it doesn't matter who you are- Yeah... 49:32 if your content is used, then it provides value, and you should reap the rewards of that value. Right. 49:38 Because, like, you know, if it's, if it becomes just a, a, you know, a negotiation, one-on-one negotiation, I'm like, I can't negoti- I mean, maybe I can, but, like, I don't think I can negotiate with OpenAI- Well-... 49:48 but, like, the New York Times can. Wikipedia is one of our partners, which we haven't- Yeah... actually announced yet. So, hey, world- Okay... Wikipedia is one of our partners, which is cool. 49:56 They today, I think, or I read about today, in the last 48 hours, did some AI deals, right? They were saying that even Wikipedia wasn't getting their phone call answered for a long time. Yeah. 50:07 Wikipedia, which is, like, the heart of where these things are being trained on, weren't getting their phone calls. Yeah. That's right. So- Yeah. So-- Oh, go ahead. Sorry. No, no. 50:15 I was just sort of saying, there is the problem of when you start getting... If Wikipedia isn't, it's not surprising that if Brian Morrissey wanted to, it, it wouldn't be as easy. Or if you're- Yeah... 50:26 a recipe blogger or if you're a creator or an influencer, that your content might not get a call saying, "Hey, do a one-off deal." 50:32 Someone has to represent that in a collective way so that there's a efficient process for getting your content into AI. Last thing is, and tell me if I'm wrong on this, Google, this is the big one. 50:46 Like, when, i-if, when Google will move to some kind of economic model, it could be this, that, that rewards, you know, publishers. Because, again, Google was the hedge-- i, it still is, the hegemon of the open web. 51:03 They determine where the traffic goes, and, and it was a really difficult job. It was difficult to balance all the different constituencies. They did fine for themselves, as well they should. 51:13 They came up with a great search engine. No offense to Ask Jeeves. No, none, none taken. All good. Okay. Uh, but, you know, they have it, and they, they ha- It, like... And that is, like, a glaring, 51:25 that's a glaring [chuckles] piece of information, I feel like. What is it gonna cost Google- So-... to- Look, it's happening already. Okay. 51:32 Ricardo Terzi, I believe his last name is pronounced, but I could have it wrong, is the head of business development for Google in Southern Europe. Yeah. He wrote just last month a post saying RSL is the future. 51:49 The head of Google business development in Europe said RSL is the future. Like, they know that this is offering them... 51:57 Now, I don't, I can't say he was saying this for himself or whether he was saying it for Google, 'cause- Yeah. Google's a big company, so, you know, I'm not sure- Google's a big company... 52:04 if Ricardo goes right to the top. [chuckles] Right. But, but what I'm saying is, is that it is making its way inside of Google to say, "Okay, if we're talking about how do we become a sustainable value add product?" 52:18 We have the value add product. We have to get to sustainability, and if we simply just take the content, it's not gonna be sustainable 'cause the content won't survive. It just won't work. 52:29 And if we try to license from everybody individually, because we don't wanna just take the content 'cause it won't survive, it costs way more than the billions of dollars to actually process it in having individual deals with every single producer and creator. 52:41 The way to solve this is the way it's been solved for hundreds, not hundreds, but over a hundred years, which is with a collective rights organizational structure. That's the way to solve it. 52:51 It efficiently gets people the content they need to be able to have a sustainable ecosystem. It's really the only way. Awesome. This has been awesome, Doug. Thank you for- Thank you, Brian... for this. 53:03 You'll have to come back next year. I was just doing a little, little Google search, and then I used Google, not Ask. Yeah. But, like, next year is the 30th- No, no... 53:09 anniversary of the start of Ask Jeeves, so, like, we should have, like, a reunion, bring back different characters. Oh my gosh. Wouldn't that be fun? [chuckles] Exactly. I would, I would love it. All right. 53:17 Thanks so much. Brian, thank you so much. This was a lot of fun. Appreciate it. [outro music]
