Subscribe
Share
Share
Embed
AI is the biggest technology shift of our lifetime. This show is about how to profit from it together. Each week I talk with the founders and CEOs closest to AI and Content, the ones figuring this out in real time. I’m also building an AI content business myself and share the lessons I learn along the way.
WHAT WE COVER
THE TITANS: How companies like OpenAI, Anthropic, Meta, and xAI are moving, and why their decisions matter.
THE INCUMBENTS: How content giants like Disney, News Corp, Universal Music Group, and Reddit are responding to AI, and what it means for creators and publishers.
THE PLAYBOOK: Real lessons on AI business models, content strategy, IP licensing, distribution, and getting paid.
ABOUT YOUR HOST: Rob Kelly has interviewed Steve Jobs and Bill Gates, helped pioneer early web content licensing, and built multiple companies with more than $100 million in total sales. His work has appeared on CNBC, CNN, TIME, and Entrepreneur.
Beyond business, every episode explores what AI means for jobs, creativity, families, and the next generation.
If you want clear thinking based on real experience in AI and media, Media and the Machine is your guide
Thanks! -Rob
What's your favorite Rupert Murdoch story?
Mark Riley:Well, first my second week during a workshop, I just presented an idea for a real estate business, which is actually still going, it's going strong. It's called Mansion Global. I presented the deck that was kind of the genesis of this idea. And it was so well received by the execs. I was literally in the lift in the elevator going to the airport when I was hauled back into the office.
Mark Riley:And they said Rupert's coming. Rupert wants to hear it. So I stood there two weeks into the job and Rupert walks in with his entourage. He sits in the front row. And I started talking to him very slowly like I'm talking to my grandfather.
Mark Riley:And his chief of staff said, no. No. No. No. Just talk normally.
Mark Riley:He's fine. Right.
Rob Kelly:Because he yeah. He was old, but at that time, but sharp. Old for maybe a business CEO, but super sharp.
Mark Riley:Yeah. He'd have been late eighties, but he was so smart. And he was like, where's that number from? Where do you get that from? What do you mean by that?
Mark Riley:And he's like, straight into the detail. So, yeah, that was one of the few times I did present to him, but he was actually quite technically savvy. He he was he was on the ball in terms of new technology and digital and the ideas we're looking at like VR and AR at the time.
Rob Kelly:I only met him once and it was barely a meeting, but you felt the electricity when he and his entourage came in. Is that what it's like on the inside too?
Mark Riley:Yeah, and I'll say he's actually quite a humble man. He doesn't sort of do the show business side of CEO. He's very humble the way he presents himself in the same way that humility can be quite terrifying. And he's also very loyal. Once you're on the inside and you're respected and you're part of the team, the company still operates like a family.
Mark Riley:So once you're on that family grouping, he's very loyal to his staff.
Rob Kelly:What was it like to work for the Murdoch family?
Mark Riley:I liked it. I thought Dow Jones was a particularly, say, forward thinking and amazingly progressive in many ways. The Newsroom at The Wall Street Journal is extraordinarily kind of neutral by definition and took pride in reporting the truth and maintaining trust. And the culture, it was quite counterintuitive in a way. It was a very friendly, very family orientated culture.
Mark Riley:And the times I met the Murdochs, it was always a pleasure to present to them and they kept us on our toes. It was fascinating to be in Midtown for eight years. Certainly felt like you were at the center of the universe for a while. It's a culture that was very grown up, but it didn't suffer fools. So you were left alone to your own devices to a certain degree, but there's nowhere to hide at the same time.
Rob Kelly:At this time, was Rupert still active in the business in Dow Jones?
Mark Riley:Very much so. I presented to him and Lachlan and James and a few others. Robert Thompson was the CEO. He was often seen in the building and, you know, a lot of respect. It was we had to walk past Fox News to get to our office.
Mark Riley:That was a challenge because they were on 1st And 2nd Floors. Culturally, that was a very different outfit and they were very much in our building. But I think the Wall Street Journal and Dow Jones was a truly separate entity.
Rob Kelly:How were they never?
Mark Riley:Well, Wall Street Journal, obviously it was a purchase. It was six or seven years before I got there. And the Wall Street Journal's politics, although the opinion pages have sort of center right and pro free markets, they aren't fully signed up to the Trump manifesto, if you like. Whereas I think if you there's no big secret, the culture at Fox was very much determined by getting the presidential choice in and then maintaining that power position. So The Wall Street Journal was a much more intellectual approach to the business community in terms of news.
Mark Riley:And as I say, the newsroom was very much driven by fact and trust, whereas our friends and colleagues at Fox might've had a slightly different philosophy.
Rob Kelly:When a big AI company and a big publisher strike a deal, I'm curious, who plays matchmaker? So when someone like Sam Altman, you know, does he call up Robert Thompson at News Corp? Does it go the reverse, or is there someone in between? How does that work?
Mark Riley:I suspect the publisher will start the conversation. They can either do it in a friendly fashion. As Robert Thompson says, you either gotta woo or sue. And I think, you know, they probably started by wooing OpenAI, saying, Hey, there's a, what's going on? And OpenAI want the quiet life.
Mark Riley:So if they see News Corp come knocking, they'll probably want to figure out an amicable relationship. That would be my hunch. I think there are other companies who've said they've been trying to get in touch with OpenAI, not getting any response without that. They then go down the
Rob Kelly:Sioux route. Large media companies?
Mark Riley:Yeah, I mean, I'm not sure what happened with the New York Times, why they kind of took their approach but it probably is a negotiating position in my mind. I would advise my clients if I was asked just kind of cut every deal you can at the moment because this is free money and it's not gonna be around forever.
Rob Kelly:Yeah, I wanted to back to the woo or sue. How do you think about that? So the New York Times sued first, sued OpenAI, then a few months later, News Corp Science, $250,000,000 deal. Just how do you think about those two different positions and then the rest of the publishing industry and what their approach should be?
Mark Riley:I think it's not as binary as it sounds. We will see. I think there is a middle ground. If you look at The Economist, the route they're taking, which I can get into. It's an interesting number that two fifty, because it's alleged two fifty.
Mark Riley:But how do you get to that number? I believe it's over five years, fifty million years. How they got to that number is a complete mystery. I mean, you could argue that they completely undersold given that they handed over the crown jewels and it's gonna potentially drive traffic away once these contents into the machines. So in my mind, it was OpenAI saying just, this is peace money.
Mark Riley:It's like a gesture just to keep them out of court. But the middle ground, the economist route is quite nuanced. It's kind of it's a love hate relationship with AI in the model. So the way The Economist is moving is like, well, we're not going to cut any deals because we don't want to give away proprietary content. And we're going to block the crawlers, but we are going to use AI internally to the maximum benefit of the business, whether that's in the newsroom or building product.
Mark Riley:So I think that's quite an interesting approach. But there's other companies like Time and I know they've been cutting every deal they possibly can with every publisher and every middleman. So it's interesting how it's gonna play out.
Rob Kelly:Is there a pattern among these? You know, when you look at The Economist and News Corp, New York Times, Washington Post, The Atlantic Time, are there two or three strategies when you're talking to media companies?
Mark Riley:I think it's more nuanced than that. If you take the fatalist approach for the creative side, it's that these models have ingested everything out there anyway, including behind the paywalls. So you can't get that content back. It's especially true of book publishers. So you may as well cream any money off the top if it's being offered, cause you're not gonna change the behavior.
Mark Riley:And then there's some more kind of defensive approach, which is either go through a pro rata or a toll bit or there's the sue and be damned approach from the New York Times who, I mean, if you wanna get philosophical for me, the legal point comes down to a kind of almost an anthropomorphic discussion around, is the model scraping and stealing? Or is it behaving like a human and reading and learning? And I think that's the nuance that the courts are gonna have to decide. So if you think that the model is almost behaving like a human, then it's just reading and learning in the same way that a human would. For me, feels like where the direction of travel is going.
Mark Riley:If you look at the most recent law cases over Getty in The UK. So I think the content creators are going to be on shaky ground. Don't say it's right, but I think it tends to be the case that technology tends to win these arguments for better or worse. So if you take the face of this argument, well, they got it all ready and they're going to win in the long term, then you may as well take any money that's on the table at the moment would be my advice.
Rob Kelly:Let me give you two scenarios. One is if you're an established media company, like some of the ones we're talking about, maybe even kind of the second tier too, because there's so many that don't get talked about and also don't get their calls answered, right, from OpenAI if you're somewhere in between. They can be sizable companies, publishing companies. If you're running one of those companies and you're sitting in the C suite, what's your strategy? Established company, been around for a while, got the normal sort of subscriptions and advertising type of business.
Rob Kelly:What approach do you take there?
Mark Riley:If they're more reliant on advertising and was serious about the zero click scenario where Google organic traffic pretty much dries up. They are royally screwed. They won't have a business model. So that's a problem, existential problem for the ad driven sites. I think if you put Business Insider in that category, they've got a fairly punchy AI strategy, which is to find efficiencies and cut deals.
Mark Riley:So I think there is a middle ground. But if you're tier two or tier three and you're not in the deal making business, there are companies you can go through to try and monetize the inference costs. ProRossa, as I say, is a good partner to sign with because they think they've got a solution to providing some kind of payment mechanism to reward any content that's been surfaced in an answer engine. So there are strategies you can do but if you're tier three or four and you're ad driven, you've got a real problem.
Rob Kelly:So if you're relying on traffic for advertising, especially then maybe you cut a deal to in some way, you get what you can for licensing your content to get through this rough patch. And if you what about if you're starting from scratch? What you've done recently, you got a couple of newsletters. Yeah. Three at least I know of.
Mark Riley:Yes, I do. So we're using our own software. We built our own software called HANA, which uses AI to create newsletters with a human in the loop, human oversight.
Rob Kelly:What's your approach to AI though, do you say? Because that's a very different calculus, right? You're starting from scratch. You don't have the Google traffic really to lose, kind of. Do you just let it be ubiquitous and go and take it as you establish your brand?
Mark Riley:Well, it's all about, if you go back to The Economist, so their philosophy is to drive direct relations with their customers. And one of the best tools for doing that is a Substack or a medium or a newsletter. And I think newsletters having a moment because they don't rely on the search engines or social media to pick it up. It's straight into the inbox.
Rob Kelly:What's your approach to folks who are seeing a huge drop in Google traffic these days in general? We touched upon this, but what's the approach you would take?
Mark Riley:Differentiate, you've got to have a brand. You've got to have really high caliber of content and quality and brand recognition, and then build up a direct, as I say, build up a direct relationship with your customers. So you're not relying on traffic that's coming through search engines or social. But if we're like being honest and we look back over the last ten, twenty years of being slaves to traffic, the metrics were all wrong. The whole kind of incentive, the business model was being wrong for so long that
Rob Kelly:How so?
Mark Riley:The idea that it just comes down to volume over quality. You could generate thousands and thousands of visits on really trashy cat videos, but that's not gonna monetize. A quality piece of investigative journalism is hardly gonna get any relative traffic, but that's completely the wrong metric. It's all about quality of traffic, where they're coming from, how much they're engaged and the value of that traffic by demographics far more important. And I think we're gonna see a swing back to that quality and then differentiate direct relationships with your customers and then just keeping them engaged and returning.
Mark Riley:And that comes down to the products. So we're back to the kind of quality over quantity argument.
Rob Kelly:Does AI change the business model at all? Does any major business model in publishing change due to AI?
Mark Riley:Radically because AI touches every part of a media company or publishing company's business. So if you think about the journey of a piece of content, an article from research to distribution consumption, the whole sausage factory is going to be reinvented with AI to make it more efficient. But at the same time, it's going to have to be quality control. But the economics, what I'm kind of obsessed with at the moment is asking my clients to move away from the efficiency argument and start thinking about new revenue streams and creating new markets and new products using AI. They've got to be quite ambitious and imaginative to do this.
Mark Riley:It's quite challenging for them. Could take them out of their comfort zones. But AI has got to be an incremental revenue business, which is far more exciting for colleagues to be building new product rather than just finding savings in the existing ones, which can scare people around headcount and job loss. So for me, we've got to be far more positive and ambitious around the business model and reinvent the business model with brand new products, with brand new revenue lines.
Rob Kelly:Let's say when we're talking in five years from now, the economics of a, call it media company, content driven company, is there something different other than subscriptions, advertising, events, like the biggies? Or is it just going to be the same types of business models and of course, a lot more automation due to AI?
Mark Riley:You missed out affiliate revenue. That's the only other one I can think of. No. I mean, I spent 10 years at Dow Jones looking at the future of news and trying to come up with innovative revenue streams. And unless you want to kind of branch out into e commerce, that really is it.
Mark Riley:So it's subscriptions, advertising, events, affiliate. And obviously we've seen it go from 10% subscription, 90 ads to the other way around, to be robust and count on those reader pace models. And I think the ones that have done that well will survive. I think they're in a strong position. You've seen the numbers coming out of the Times in London and New York Times and Wall Street Journal that they've got very, very robust business models now.
Mark Riley:So if you add in automation from AI, that's gonna see your EBITDA get better and better as long as they don't damage the product. But the bit they've messed up is the distribution model and the reliance on social and mess in Facebook and Instagram. Those models are so unreliable that coming back to my previous point, you gotta build your own reliable direct distribution models. So you're not wholly reliant on these third parties that have completely different values and incentives to the publishers.
Rob Kelly:So the main economics, probably not some major new source of revenue as big as some of the ones we're talking about, but you see a flip in subscriptions and advertising, which makes sense. You want the direct relationship. What about are we heading to a time where folks are gonna start their day of consuming content via AI? In other words, will I be reading The Wall Street Journal in OpenAI or an OpenAI app as opposed to opening up the print paper, which I'm old fashioned, I still do, or reading it online? Are we gonna be reading content starting our day with AI?
Mark Riley:I think you might be listening to content on AI. I think the listening use cases is getting very interesting in terms of transcription tools and text to voice. So you no longer have to have the Wall Street Journal open in front of you. You can be listening to it on your commute, on your drive. Personalisation is the big watchword of 2025.
Mark Riley:It's like, how do I deliver news that's relevant and timely in the content format that the customer wants it? So some people wanna ingest The Wall Street Journal as is, some people might prefer to have it in the form of an Excel spreadsheet and other people might want it as a TikTok video. It doesn't really matter. That is fundamentally sacrilegious to the editors of old because they take great pride in their copy and their presentation. But you've got to break down these shibboleths and move on and understand personalization.
Mark Riley:You might get an OpenAI chat GPT feed in the morning that's kind of brought you the best of the content from the five publications that you follow in the same way that Apple News does now or Google News. So it's going to become more atomized but more personalized at the same time. And I think publishers have to be open to that model, which does come back to the problem of attribution, which is where we're talking about companies like pro rata. So if you're getting rather than having a boiled egg for breakfast, you're getting scrambled egg from five different publishers. How do you figure out who's contributed to the scrambled egg?
Mark Riley:That's an interesting challenge at the moment.
Rob Kelly:How are those companies doing? The toll bits of the world, pro radas, are they adding substantial value to any publishers?
Mark Riley:Jury's out. I think too soon to say. Hats off to Pro Rasa for building so many partnerships, whether or not they have the ability to redistribute revenue. Two questions come to mind on that. The first one is how do you split up?
Mark Riley:As I say, how do you attribute? Is it just done on citations or source referencing? And the other question is what revenue you are splitting up? So where's the revenue in the top of the funnel? Is it user subscription to Pro Rasa or is it advertising revenue around the pro rata gateway that's being split up in the same way that Perplexity is trying to manage?
Mark Riley:So there's a lot of open questions and I wish them well. I think they're well intended and I think they're doing something worthy. But when it comes down to the nuts and bolts and delivery, I've still got quite a bit of British sort of snark around it.
Rob Kelly:We'll get back to that. Will AI send traffic to publishers in the way that Google did?
Mark Riley:No, opposite, absolutely opposite. It's catastrophic. We're already seeing clients seeing or publishers seeing 80% drop in Google organic. And it's an injustice if you think about it cause the content they're serving up and their summary answers are almost certainly drawn from the publishers themselves. So it's gotta be figured out.
Mark Riley:I know Perplexity are trying to bring the publishers back on-site by offering a big chunk of the subscription fee. But at the moment that's going to be pennies. I think the best way to keep the publishers on-site if I was a frontier model would be somehow drive subscribers back to the publishers. Cause that's the main business they're after and a huge lifetime value in a subscriber. So I think the AI is smart enough to know through data and seeing the user come through and what questions they're asking and where they end up, they can filter out the most likely to convert.
Mark Riley:And if they can then send the publishers high converting quality subscriber traffic, I think that's an interesting business model that they should be looking at.
Rob Kelly:Interesting. So the idea, I mean, do you think we'll get to a spot where the LLMs are using citations and those citations maybe even are links to subscribe to The Wall Street Journal, subscribe to whoever the content came from. What's the chance we can get to that spot?
Mark Riley:I think that technology is there already. Think it's just getting these people in a room to have a grown up conversation rather than just trying to sue the head out of each other. I think there is an amicable way forward, but it is without a doubt the zero click scenario is a very serious one. And it's the smaller publishers that will wither on the vine. And then you've got the kind of if these models kill off the smaller publishers, they're not gonna have any content left to ingest.
Mark Riley:So they do have an ulterior motive to keep these businesses going. And by the way, do not think they can do journalism. These West Coast frontier AI companies do not understand journalism. They don't understand media. They don't understand publishing.
Mark Riley:It's been proven over and over again. There's complete cultural mismatch between the two ways of working. And they need to sustain quality journalism to give them the quality content they need to feed back into their machines. So it should be a symbiotic relationship, but they've got to figure out the value exchange.
Rob Kelly:Maybe you can give some examples when I want to get your quick sort of hot take on the LLMs. I'll just list through them and maybe just give your one liner on how they are with content and and media because they're not all equally bad. Right? I'm sure there's some, shining moments there somewhere. But let me just rattle through them, and you just tell me how how you think they're interacting with content as a business.
Rob Kelly:So why don't we start with OpenAI? Diplomatic.
Mark Riley:They're putting money on the table. They're willing to cut deals. They understand the ecosystem. They'll see how long it lasts, but they understand that probably they got Microsoft breathing down their necks. They just don't wanna have 30 lawsuits on their desk in the next three months.
Mark Riley:They're just playing a diplomatic game. How about Anthropic, Claude? Well, they've just compensated book publishers to the tune of 1,500,000,000.0 again to avoid any further damages. So they've come around to the idea. They scared off by fully putting their hands up and saying, yeah, we've illegally ingested all these published books.
Mark Riley:So here's some peace offering. But these guys, they're playing fast and loose. So in a way it's good that you have a good cop, bad cop relationship with these people. So some of the publishers are suing and the others are wooing and that seems to be getting the effect they want in the short term.
Rob Kelly:How about Google and Gemini?
Mark Riley:Well, I mean, no one dares sue Google until this year cause they just don't want to rock the boat. They do have one lawsuit against them. They're obviously the ones that are going to have the most damage in the sense if they freeze up the organic traffic. It's an interesting one. I mean, it would take a lot of bulls for a conglomerate of publishers to take on Google, which they have done in the past, especially in places like Australia.
Mark Riley:And again, News Corp have been through the machine. So they are beatable, but they control the internet. So good luck.
Rob Kelly:How about xAI, next.com, Grok?
Mark Riley:Honestly, I mean, they've got enough data to work with because they're ingesting the Twitter feed and that's their principal training data say. And that the percent of queries that are going through Grok are fairly irrelevant at the moment. So it's less of a worry but again, good luck suing Twitter hacks. Deep pockets.
Rob Kelly:Yeah. And perplexity.
Mark Riley:Okay, so their argument is that there's a misunderstanding between what is a bot and what is an agent and what is a crawl. And they believe that if you're using their browser, Comet, and you're instructing an agent to go off and do some research for you, you're effectively treating it like a human assistant. And therefore that agent has every right to go to every website and crawl and discover and scrape and bring the information back. Again that's going to have to go through the law courts. I think they need to win that argument or else they will be in trouble.
Mark Riley:They have been caught out. They've been accused by News Corp and Dow Jones and others of and Reddit of caching paywall content. And there's evidence of bad behaviour by their bots ignoring TXT texts and going behind paywalls when they've been asked nicely Nazi. So they're in a very grey area. They have built a thing of beauty.
Mark Riley:I think they're amazing product people. I think they've built a phenomenal product. I love it myself. I find it extremely useful. I'm probably using it way more than I use Google now.
Mark Riley:So in terms of product, I think they're way ahead. But in terms of behavior, they're gonna have to figure out how they compensate. And now they are bringing out this publisher partnership programme, But I'm not yet convinced that the numbers stack up meaningfully to compensate the publishers. But we'll see. Conde Nast have come on board with them, Washington Post, a few others.
Mark Riley:But yeah, I wish them luck because I think that enough users are addicted to their product, which is phenomenal, then they should win through but they've gotta figure out this relationship.
Rob Kelly:You mentioned agents before. How will agents interact with content?
Mark Riley:Well, interesting case in point that Amazon's suing Perplexity's comment to stop their agents going shopping for them.
Rob Kelly:What's your take on that?
Mark Riley:It's again, is a negotiating position, but the big kind of flaw with agents is they don't read advertising. So Amazon is essentially, as far as amazon.com is concerned, essentially an ad driven business as much as a retail business. And if the agents go and shopping, they're not gonna see all the superfluous ads that's driving you to put other stuff in your basket. So that's a problem. I don't know if you can train the agents to get influenced by the ad experience, but that is a problem for Amazon.
Mark Riley:So they clearly put their marker down. So if you use deep research now on ChatGPT, it's essentially an agentic task. You can see the agents or you can actually go into agent mode on ChatGPT and you can see them opening browsers and doing your work for you. So they are getting very good at deep research, putting reports together, analysing different sources. And that's a very good agentic use case.
Mark Riley:But again, as I say, they're interacting with the websites as a human would, especially if they're not reading ads and especially if they're dodging paywalls.
Rob Kelly:Yeah. I'm still having to ask each time for, you know, I always want the citations. I want the sources. At least I use ChatGPT for my deep research and they just don't seem to on the side of linking to where they gave me information from.
Mark Riley:What you can do is put the content into NotebookLM and then ask NotebookLM to go and verify the content and come back with the citations. It's one workaround.
Rob Kelly:Will AI cause the ad industry to get bigger, smaller, about the same?
Mark Riley:It depends how you're defining the industry. If you look at the agency models, so WPP, for example, Omnicom, I think they're really struggling to keep up with this world where content or advertorial or adverts can be created with minimum effort now. And the whole need for kind of armies of content creators, videographers and buyers is gonna be obsolete. So I think the entire agency model needs to be reinvented. I think there's an argument that AI will help ads become more effective and personalised and relevant and contextual.
Mark Riley:If the AI can match the ads to the content to the user, I think you're going to see ads become even more powerful and convert better. But yeah, back to the bot argument, if 80% of web traffic is gonna be bots, then that's a problem.
Rob Kelly:Do you see, I know excluding the bot scenario, do you see affiliate ads you mentioned earlier is getting a boost as a result of AI?
Mark Riley:Yeah, I hadn't really thought about it but the affiliate ads is purely about math or maths as we say. And I think it's quite easy for AI to figure out how to optimize conversions from funnel to advertising to product to completion. So I think that'll help. But if you look at the main use case for advertising in AI right now is Facebook advertising. And the reason why you're seeing a lot of increase in Facebook revenue, of course, on course, is because the amount of AI they've embedded into Facebook ad targeting.
Mark Riley:They're probably market leaders at this. They've been at it a while, but it's based on the amount of data they have in hand because they know so much about us, you, the user, that they can target Instagram and Facebook ads extremely well.
Rob Kelly:Is it inevitable that the big LLMs will sell ads?
Mark Riley:I'll ask Sam Omer if their revenues are stressed or they're not hitting targets. They've obviously got an enormous amount of ad inventory they can turn on overnight. The problem they've got is that it can't corrupt the integrity of the answers that they're generating. So at the moment it's a really nice experience to go to any of the LLMs or Perplexity and you get a pretty pure answer that's responding to your query without bias. But as soon as you're steering, which is the best pair of sneakers to buy for a 10 ks running programme?
Mark Riley:At the moment, it's going to do some pretty good content reviews, product reviews. But as soon as the advertisers are on board, you're not gonna trust the results half as much. So we'll see. I think there'll be a huge resistance internally to going this route. But if all these checks that they promised to sign for GPUs isn't kind of measuring out and they need to double their revenue overnight, they've got one avenue that might have to go down that.
Rob Kelly:So we talked about classifieds in our last conversation. I've got some more detailed questions, but relate to AI and its impact on it. But can you just start off with giving a kind of a primer on the classifieds business industry? You worked for The UK version of Craigslist, right? Gumtree?
Mark Riley:Yeah. Well, if you look at the history of classifieds, so the classifieds and especially real estate or property, as we say, was the main backbone of local media publishing newspapers, let's say. Newspapers, turn of the century up until the sort of 2010s. And they lost the business overnight. They were asleep at the wheel.
Mark Riley:They allowed the Craigslist and then the Gumtrees and the Rightmoves and the move.coms to set up shop and go down the real estate niche. And they stole that revenue from under their noses. So it's a highly, highly lucrative game, especially luxury real estate, being able to advertise properties to high net worth. So what you're seeing now over the last ten years is that the publishers trying to win that business back by acquisition. So you saw how News Corp bought move.com back into the business.
Mark Riley:REA is one of the leading revenue generators. So it's kind of gone full circle. Where it goes from here is interesting if you put an AI layer over classifiers. And this is a hypothetical question, is what if you train a model to be able to find a classified ad that's highly relevant to your search criteria? Does it do the same to the platform business, the Brightmoves and the REAs that OpenAI are doing to the publishers?
Mark Riley:In other words, hypothetically, could you scrape all the classifiers that are out there? So I'm looking for a particular property in a particular city or town, and I've got very tight criteria. Would AI be able to surface that ad without me having to go to the destination platforms? I don't know, but that would be an interesting take. It might get better when a Gentex search takes place and it learns how to navigate into the websites.
Mark Riley:Actually execute a search on your behalf, that'd be interesting. But, you know, there's been classified aggregators around for a while that just present another layer on top and then send traffic back. But if you're thinking of AI as a personal assistant and shopping as one of those, then browsing classified ads for second hand clothing or real estate or jobs is going to be an interesting task for them to do, which will fundamentally undermine existing classified business model.
Rob Kelly:I'd have to go back, but I'm a little bit of a student of internet and tech. I'm trying to think of the waves and how they impacted classifieds. So you had the Internet come along and that enabled Craigslist to take a bunch of the classifieds business from print. Right? What are the other big waves?
Rob Kelly:So I guess you had mobile come along and with GPS that enabled Zillow to really become a killer app. Right? Because you could just look up, you know, wherever you were. Right? Go look at houses for sale or for rent.
Rob Kelly:Yeah. Wouldn't AI just enable some new innovation in real estate and classifieds in general?
Mark Riley:I think so. Ultimately, would you be able to disintermediate the entire platform structure? If I'm selling a house, can AI match me with a buyer, fits my criteria, and vice versa, without any kind of intervening layers? So that would completely undermine the Zillow business model. Interesting question.
Mark Riley:I think you'd need some kind of database. That's what the platforms are providing at the moment. It's kind of an index searchable database. So if you took that away, I think the models would struggle. It comes back to the kind of crawling, scraping, stealing debate.
Mark Riley:If you could train an agent, just go off and do your house search for you and then connect you with five or six properties, chances are you would still have to visit the website to get to the seller. But would AI be able to connect you straight to the seller? That would be bad news for the Zillows. And the property is definitely, it may just be what we call property poor and it might just be people browsing for high end. But my background obviously built Mansion Global for News Corp for Dow Jones.
Mark Riley:So I've got a kind of background in that area. I think there is revenue to be made in super high end real estate if that's the way you want to sustain a newsletter.
Rob Kelly:How big did the Wall Street Journal mansion business get?
Mark Riley:I don't know if I could say. I don't think they break it out in their results, but it was doing over 10,000,000 annual recurring when I left and that was five years ago. And it was growing nicely. So it was a good business.
Rob Kelly:And the Wall Street Journal doesn't launch new sections very often, right?
Mark Riley:Well, it was weirdly a digital extension of the mansion section of the Saturday edition of the Wall Street Journal. It was leaning on the brand but just turning it into a global product rather than a local product. Was mansion considered a classifieds business? Would you call it that? There's some good content in there.
Mark Riley:Property produces some very good content around where to live and renovations and property of the week. But ultimately the money is coming from the Christie's and the Savills and the Sotheby's who were looking to reach a global audience, and this was a great vehicle for them.
Rob Kelly:Right. So Mansion has human generated content. You see these big 25,000,000 homes, and they say it in the article, but written by a human. And then you've got the Sotheby's real estate ad or one of the realtors paying for advertising.
Mark Riley:Yeah. It was very bad for my self esteem having to look at these $25,000,000 properties all day. Yeah. Yeah.
Rob Kelly:I hear you. What I do is clip out, like, you know, the one little section that like a fireplace that I can afford and and give it to my wife and say, this would be nice. I hear you, though. So back to Hannah. So I noticed you've got some advertisers in there too.
Rob Kelly:You've got I I just noticed advertisers popping up like realtors, like Seville's, I guess, and Rightmove and Ideal Home and UK Property Forums. Are they advertisers? Are they paying you for appearing in the newsletter?
Mark Riley:Not yet. No. We're just building the audience right now, but we're being quite diligent to make sure we have more than one advertiser on there. And then once we can measure the traffic and a click through, we can then go, you know, we've got to build some trust and continuity and keep turning up. So I think Bristol's on edition 60 something, Oxford's on 40 something.
Mark Riley:And I think once we've got some continuity and audience, we can then go to these brokers and say, Look, we've got your audience, we've got your interests, let's try and cut a deal. But no, we're still very much in the growing stage.
Rob Kelly:I know you've got events in your background, in your career. In this new world of AI, how do you see real life events in terms of media business being impacted?
Mark Riley:The number one way, I mean, it'll come back to that if we go to zero click scenario, number one way to build relationships with your audience is directs. And the number one direct way is events. I think it's a kind of irony or paradox that the more digital our lives become, the more we crave human contact in the real world, whether that's technology events or sporting events or music events or business events, It's a key revenue driver. And I think that my time at Wall Street Journal, they were very good at it. We had the Future of Everything event and meat packing, which is one of the best events I've been to.
Mark Riley:And it just creates a real buzz. So it's like untouchable. I mean, there is a role for AI in event planning and event management and running the data behind an event but ultimately an event is a very non AI event. It's a place to escape AI and actually interact with a human on stage. There's absolutely, as long as they're not avatar, it's gonna give us some bit of humanity back.
Rob Kelly:So you've worked in both The US and UK, you're born in UK, from the accent. I'm fascinated by kind of the resurgence of The UK in tech due to AI. Matheson is the middle name of Alan Turing. I didn't know that until you told me, but I think that's fun. Is I could tell you got some UK pride.
Rob Kelly:You got Google DeepMind, DeepMind first when they were without Google starting there, bunch of exciting AI companies. You got Oxford, you got Cambridge, both with AI labs. Right? Very reputable. Yeah.
Rob Kelly:Just tell me about life in The UK and is it a resurgence of UK pride? Because, you know, us Americans used to always make fun of anyone outside of The US for being a couple of years behind on tech. And now you got, you know, it's tough to argue that anyone's more important than DeepMind with the formation of the new world of AI. OpenAI, obviously, but I mean, the DeepMind folks were the pioneers. Yeah.
Rob Kelly:How do you feel about all that?
Mark Riley:Well, you've it in the sense that our strengths are our university and research, and we're phenomenally good at ideas. Cue the dog.
Rob Kelly:He or she agrees.
Mark Riley:We're less good at the growth, mainly due to, I think the funding structure and the lack of capital and the lack of risk appetite in this country compared to yours. So we do see the most successful startups will go to the Valley, to Sandy Road to look for funding. And that point they will then be asked to move to California and then they've gone out of the ecosystem.
Rob Kelly:And they usually move except for DeepMind.
Mark Riley:DeepMind have stayed put. I mean, we've got some rock stars in the top 10, top 20 apps at the moment, like 11 Labs are doing very well. Fix is on the app, a couple of others. And I think we are creating our own little sort of Silicon epicentres around London and Cambridge and Oxford. Britain and Europe are way behind, both America and China for obvious reasons.
Mark Riley:So but just to correct you, the internet was quote unquote invented by a Brits, Tim Berners Lee. So we'll take that one as a win. As you say, DeepMind has been hugely significant in powering early deep neural nets and going forward. I would say we're probably number three or four, both in terms of research papers and startups. So I think Britain is punching above its weight, but way, way behind the pack.
Rob Kelly:What's the BBC gonna do, AI wise, do you think? Got some good content?
Mark Riley:Yeah. I mean, they haven't let anyone near their content in terms of using it for training or inference. If you think that even then, their natural history content going back fifty years is an absolute gold mine of trainable data.
Rob Kelly:Isn't that headquartered in Bristol? Yeah, it is. But overall, do you think the BBC will do AI wise?
Mark Riley:Well, they might have to change their tune because they're under a lot of commercial pressure right now but they're not commercially minded in the same way as a publisher would be in The US. So there's no prerogatives kind of offer their content away for either training or inference. But I know they've been making preparations in terms of embeddings and tagging and getting it all in order. So we'll see where they go. But you're right, they are sitting on an absolute gold mine.
Rob Kelly:All right. I've got some final humanitarian questions. You've described yourself as an AI optimist, accelerationist. Can you say more about that?
Mark Riley:Yeah, I think there's a tendency to jump to the kind of existential threat of AI is going to eradicate humanity as a species debate, which I love that debate and I love engaging in it and I have my own concerns and worries but I'm an accelerationist in the argument that I'm not a kind of crazy neoliberal techno enthusiast that thinks it's gonna trump all other needs in humanity. It's much more a question of when I say acceleration is, I mean, it's a race, both as a society, a community, a business and a country to stay ahead in this game.
Rob Kelly:What do you think of universal basic income?
Mark Riley:I think they need to start doing the math because this is going to become very real very soon and much sooner than people think. What I worry about UBI is not so much the financials of it, it's a question of purpose. And I think humans need purpose and they need the dignity of work and they need to go out every day and achieve something meaningful involving people. So it's all very well putting people on a reasonably generous sustainable income. But I think without purpose, we're heading for trouble.
Rob Kelly:What are you telling the young kids in your life about changes from this world of AI?
Mark Riley:Well, it's a two pronged approach. One is to engage with it, be good at it, be very familiar with the tools. And it's interesting that people entering the workplace under 25 are completely at home and au fait with that. And that gives them a superpower of sorts, because the older colleagues will be looking to them to teach them how to use it and give them confidence. But at the same time, there's a question you've got to be, I'm a great believer in lifelong learning and adaptability and being able to switch paths very quickly.
Mark Riley:And just to steal a line from Scott Galloway, it's like, what should you be telling your kids to be good at? And his answer is, they've gotta be good storytellers. And I think there's a lot of weight in that argument, in that response, which means be good at presenting, be good at conveying your argument and be good at communicating.
Rob Kelly:What if in a scenario where AI is doing all the work that we're used to doing and you had endless time, what would you do with all your time?
Mark Riley:Well, I've got this. Have you heard of the fit dog theory? No. So there's a lawyer who's using AI and he got an hour or two back a day and so his Labrador got a second walk each day. So he had a very fit Labrador.
Mark Riley:So that's my fit dog theory of AI that people will find leisure time and fulfilling leisure time.
Rob Kelly:And will you or have you already, or will you create an AI avatar of yourself for your family, friends and business so that those close to you can have conversations with you after you've passed away.
Mark Riley:I think bereavement, I'm not saying that I'm not being talking about me but I think bereavement is a very natural healing process and it allows us to move on. But if we create everlasting avatars, we're never gonna move on. And that fills me with dread. I think that's horribly dystopian. Wouldn't go near it at all.
Mark Riley:It would discourage anyone from even contemplating it.
Rob Kelly:Thanks for investing all the time, Mark.
Mark Riley:Thank you, Rob. It's been a pleasure. Great chatting to you. Thank you so much.