Skip to playerSkip to main content
  • 1 day ago
Transcript
00:00This is a story about big changes on tiny screens.
00:04Since its inception, the Internet has become where we turn for content,
00:08content we mostly access through search, and that's mostly free.
00:12But, of course, it's not really free.
00:15When we searched on platforms like Google, we were directed to web pages
00:19created by publishers who got paid for all those ads we spent time looking at.
00:23Now all that is changing, and as AI becomes the new gateway to the web,
00:28it's transforming the economics of the Internet.
00:35We believe the high-water mark for Internet traffic that was sent out
00:40from social media and search and AI tools, which were quite nascent at the time, was 2022.
00:46Essentially, over the last four years, those numbers have been dropping.
00:51Rand Fishkin is an author and entrepreneur who specializes in online search.
00:57The main way that most of us interact with the web.
01:01I do a lot of research with clickstream data, which essentially looks at devices
01:06and what they visit and don't visit.
01:08Whereas in 2011, Google was sending north of 70% of all searches to a third-party website,
01:17right, a website not owner-controlled by Google.
01:19And today that number is in the mid-40s and falling.
01:26That clearly indicates this zero-click trend.
01:30Zero-click.
01:31It's a term for an online search result that doesn't lead to any other web pages.
01:36That can be good news for users if their question has been answered quickly and correctly.
01:41No need to spend time scrolling through all that text and pages of ads.
01:45But if you are a publisher counting on that click to bring traffic to your website and to sell those
01:51ads,
01:51it's very bad news, which means the creators of all that web content have to rethink the game.
01:58Long before AI, Google started disintermediating content providers by answering questions right in the search results.
02:05You search for the MET score, the score's right at the top.
02:08And AI just enhanced their ability to do this.
02:13For an idea of just how big this change is for Google itself,
02:17here's the company's head of search talking about incorporating AI just last month.
02:22Now we're entering the next chapter of Google search.
02:25Where incredible AI features aren't just in search,
02:28Google search is AI search through and through.
02:32Now this is the biggest upgrade to our iconic search box since its debut, over 25 years ago.
02:36The problem is that most people don't click on the links.
02:41Caitlin Petri is author of All the News That's Fit to Click,
02:45and a professor at Rutgers who studies the way algorithms shape online media.
02:51News sites that were counting on referrals from Google for their traffic, for their ad revenue,
02:58to draw maybe subscribers who might come to the site, get interested, and then subscribe,
03:02are now seeing, in some cases, very dramatic decreases in the traffic that they're getting from Google.
03:08The move from print to the internet, many people think, actually eliminated some news sources.
03:14A lot of local newspapers, for example, were eliminated,
03:17partly because of the business model challenge, the revenue challenge.
03:20Do we face a risk of that with AI that we will eliminate even further sources?
03:25I think we do. I think it's a major concern.
03:28It was really revealing.
03:30Perhaps nobody in the world has a better view of these changes
03:33and what they mean for online content creators than Neil Vogel.
03:37What happened was our audience, right, and we reach half of America each month,
03:44our audience on the internet was 75% from Google search.
03:4920, 22, maybe a little bit more, maybe a little bit less, but ballpark that.
03:52That is now 25% of our audience.
03:56Vogel is CEO of People, Inc., the largest online and print publisher in America.
04:01It reaches 175 million people each month,
04:05with dozens of publications whose names you know well.
04:09Better Homes and Gardens, People, Entertainment Weekly, Travel and Leisure, Food and Wine.
04:14And we put all of these brands together because we believed we understood
04:19how to run these classic heritage brands on the internet.
04:23And we've been proven right.
04:25Although Vogel has had a front row seat to the collapse in search traffic that AI has ushered in,
04:30he also told us something surprising.
04:32His business is booming.
04:35We're very profitable, and we're defying the narrative around publishing
04:39that as publishers can't succeed in this world, and we're succeeding.
04:43We feel really good and really optimistic about the future of media.
04:46Back in the day, before there even was an internet, believe it or not, there was a time.
04:50If you look at a publisher, newspaper, magazine, their revenue came from two sources,
04:55subscription and advertising.
04:57What's your revenue model?
04:58If you look at the internet, that is where we thrive.
05:00That is our bread and butter.
05:0290% of our profitability comes from our internet and our online presence.
05:07That is very much advertising, but it's also lots of other things.
05:10Those other things include deals and branded products.
05:14But, Vogel says, they've also got a new revenue stream.
05:18We license content to LLMs and AI engines.
05:22We sat with Sam Altman at OpenAI right before they could launch the first commercial LLM.
05:29It was the fall of 2022.
05:32And we saw this LLM, and we saw what it could do, and we were like, uh-oh, the world
05:38is going to be different from here.
05:39And there's moments in your life where you know the world is going to be different.
05:41You have your first kid or the Philadelphia Eagles win the Super Bowl, and you know that your world is
05:46going to be different from there.
05:47This was one of those moments.
05:48This was like the business version of that moment.
05:50And we took a look at this, and our first reaction is, thank God we have brands.
05:55Because what we knew was this was going to be a new paradigm in media, and we knew there were
06:00going to be risks, and there were going to be opportunities.
06:02And we made the bet and had the thesis that says, if you have these intangible assets that really, really
06:09mean something to people, you have something that is going to increase in value in an artificial era.
06:14And starting in 22 and 23, we started to build other ways to distribute our brands.
06:19TikTok, Instagram, YouTube, our own emails, product licensing, all of these things we did.
06:25And cut to today, five years later, we've made 10 straight quarters on Wall Street.
06:30We're growing in an area where most publishers are struggling.
06:34And we did it because we have these amazing brands that can live in all these places.
06:39And our aggregate audience size is multiples what it was when we started here.
06:44You mentioned that you do have licensing arrangements with some of the LLMs, large language models.
06:49How does that work for you?
06:51Is it real money?
06:52So it's real money.
06:53It's material for us.
06:55I think we made a decision, maybe it's right, maybe it's wrong.
07:00So far, I think it's right, that we wanted a seat at the table.
07:05The important thing to understand and realize is AI needs three things.
07:08It needs power, which has been a lot of talk about power.
07:11It needs a model.
07:13It's been a lot of talk about models.
07:14And it needs inputs.
07:15We are the inputs.
07:17We are the training of the models.
07:18Now, we believe, and I think if you talk to any of these AI professionals believe, the Internet's been crawled.
07:24All the data in the world is already in the models.
07:27So the value should accrue to those of us that are still making substantial amounts of new things that are
07:33needed for the search use case, which is a very big use case, and are needed for all the corporate
07:38clients that want to use LLMs for their stuff.
07:40So it could be something simple as we have 15 new recipes with avocados in them, and avocados are important,
07:45to I can tell you every single thing Kim Kardashian did yesterday.
07:48But these are things that are very valuable, these pieces of data.
07:51The other thing we did, which is really important here, is part of having a seat at the table is
07:59we wanted to control our own destiny.
08:02Last summer in July, we made a decision, and we partnered with a group called Cloudflare to block every AI
08:09crawler you could possibly block.
08:11So you cannot get access to our content unless you pay us for it.
08:13And given your model to license the training using your content to an LLM, it's basically found money, right?
08:23It's not taking away from your other business because that's based on brand?
08:27I mean, so it's interesting.
08:29Is it – it's funny.
08:30We have this conversation all the time.
08:32Like, and the answer we always come back to is cash is cash.
08:36It would be amazing if we could make money the way we made money three or four years ago.
08:40It would be super fun.
08:41I'd have a lot more free time, but I don't.
08:43And we have to work hard.
08:44So it is what it is.
08:47For a publisher with strong, well-established brands, even the uncertain future with AI may be bright.
08:53But for smaller publishers, and for the web as a whole, the outlook may not be as encouraging.
09:00Potentially, in the longer run.
09:03What could this mean for the Internet itself?
09:05Because the Internet really has thrived on very rich production of content.
09:10If content providers don't get paid for that, sooner or later, there's not going to be as much content.
09:16Yeah, this death spiral that you're describing has been something that people have theorized over the last three or four
09:22years with no discernible solution.
09:26I'm going to tell you what I think is the most likely scenario, but it's also the most tragic one,
09:31which is that in a few years, AI tools find themselves unable to get enough content to train on, whether
09:39that's video or text or audio, long form, short form.
09:42And they decide to build their own publishing departments.
09:45And in years to come, a decade, two decades from now, it would not surprise me if OpenAI or Anthropic
09:54looked a lot like TikTok, where they have a creator fund, and they pay creators a certain amount based on
10:01the amount of traffic, page views, and value to the training set that these creators provide.
10:06And so they're contract employing these creators, and we get an AI-shaped Internet, right, an Internet that is built
10:16not only for, but by AI tools.
10:21That seems quite probable.
10:23I think that that is something that if I worked at an AI company, I would really be thinking about,
10:30because if the traffic drops so significantly that we see even more jobs in journalism be lost, we see even
10:36more, especially local outlets, but a lot of news outlets shutter themselves because they can't stay afloat, that diminishes the
10:43amount of content that these LLMs can train on, too, right?
10:46So this is something that I think we should all be concerned about, not just journalists, but also those who
10:53work at AI companies, anyone who cares about the future of democracy should be thinking about this right now.
10:57Tell me how broadly, apart from people, beyond people, how do you think generative AI is going to change the
11:05nature of the Internet, the content on the Internet?
11:07So we can end up being fairly myopic around here and looking at our own use case.
11:13Yes, I am, and I have to be, I'm like a crazy optimistic person, and I'm very optimistic about AI
11:23for us, right?
11:23The risk for AI for us is already out.
11:25It's already killed our biggest piece of traffic from five years ago, but we've built something even better over here
11:31that we use AI to do more of.
11:36Like we are today, we make about 50% more content than we made when this whole process started for
11:43roughly the same cost, and humans make every single thing you read and see, because we've been able to use
11:49AI in the background to streamline so many processes.
11:52If you look at it as an opportunity, it's incredible for us, and we have no choice but to look
11:58at it that way.
11:58If we look at it any other way, it's too scary, and it's sort of like riding a bike.
12:02Don't look down, look towards the horizon, and you'll be fine.
Comments

Recommended