Paul Ford: Hi, I’m Paul Ford.

Rich Ziade: And I’m Rich Ziade.

Paul: And this is The Abord Podcast, the podcast about how AI is changing the world of software. For better or for worse. Sometimes for better. It’s brought to you by Aboard, aboard.com. Check it out. We’ll mark it a little more later, but for right now, we have some really important stuff to talk about. Rich, have you ever heard the letters “GEO” before?

Rich: G-E…Oh no!

Paul: Exactly. Let’s get into it and play that theme song.

[intro music]

Paul: There is a situation on the internet right now. You got a billion people, give or take, a couple hundred million, using AI on the regular.

Paul: Yep.

Paul: Let’s talk about how it used to work for a second. Name a good cuisine.

Rich: Vietnamese. Love Vietnamese food.

Paul: You’re a Vietnamese restaurant in Brooklyn. Zip code 11215.

Rich: Yep.

Paul: You put up a nice website.

Rich: Cool.

Paul: You put your menu on. What else you put on?

Rich: Links to Resy, if you want a reservation. Pictures of the food.

Paul: About us.

Rich: About us. Location.

Paul: Menu.

Rich: Menu. Yes. All those things.

Paul: Okay. You put all that up. What is the implied transaction you’re going to have with Google, as a website publisher?

Rich: Well, I mean, the premise behind all of it is that people are, 95% of the time, are going to search on Google to find you.

Paul: That’s right.

Rich: And so Google has sort of laid out ground rules around how to, like, structure your site, how to, like, signal to it, so that it can index your site, and so that when someone types in, “Vietnamese near me 11215,” you hope that your site and your information shows up high up on the list.

Paul: This is the foundational contract of the commercial web, in a funny way. Like, it’s not a legal contract.

Rich: No, it’s why—

Paul: But it’s understood.

Rich: That’s right. It’s why Google became a behemoth, is it became an intermediary to getting information.

Paul: That’s not really how Google makes money.

Rich: No, that’s how you capture intent.

Paul: Right.

Rich: And audience.

Paul: Google knew that people wanted to eat Vietnamese food, and that they would type in “Vietnamese Brooklyn.”

Rich: Intent fuels economies, right? And so they put themselves in the middle.

Paul: So now how does Google make money, given that people want to do that?

Rich: You can buy ads.

Paul: So that Vietnamese restaurant could buy an ad.

Rich: Not only could they buy an ad, they could buy an ad, if you search for their competitor down the street.

Paul: Right.

Rich: Like, you can actually bump yourself up so that you’re the first result when someone types “Joey’s Vietnamese” down the street.

Paul: I mean, you see it all the time. You search for, like, Pipedrive and Salesforce comes up, if you’re looking for a CRM.

Rich: I mean, they put themselves as the middleman for adversarial marketing, which has been a huge source of success for them.

Paul: And then Google also will help you run ads on your website, if you’re a publisher.

Rich: There’s all sorts of ways.

Paul: So there is this enormous ecosystem where that intent eventually would get somebody to a website.

Rich: Yeah.

Paul: And that website was sort of the golden thing. It was the thing we were all building and working on.

Rich: Yes.

Paul: And we would try to make the best possible Vietnamese website. Some people might really try to hack it. They might try to make a bunch of fake content or fake links or whatever, or they might be legit. And Google would sort of sit there in the middle and say, “Okay, I will show it to you, but hold on just one minute.”

Rich: Yeah.

Paul: “I got to make a couple fractions of a penny, and now I am worth several trillion dollars.” Okay? So that’s kind of the fundamental.

Rich: Here we are.

Paul: And that’s the whole web. Like, Reddit is that. Why does Reddit exist? What can you do with—how does Reddit make money?

Rich: Reddit is interesting in that Reddit has a search feature, so you can search Reddit, but you’re not as intent-driven. Where is there intent? There is intent in communities. Communities gather around topics that matter to them, whether it be coding or music or hobbies or whatever it may be. And it turns out those communities, some of them are really big, and you can situate yourself within those communities as an advertiser.

Paul: That’s right. And those communities also generate a lot of value because they generate a lot of good, opinionated content about products.

Rich: Yes.

Paul: But basically the model has been we make a lot of content—somebody makes it, either users, or the person who is the restaurant owner.

Rich: Yes.

Paul: And it goes on the web, and it’s supposed to be pretty useful, at some level. And then there’s an advertising ecosystem built on top of it. I mean, I’m glossing over, like, YouTube here. Like, there’s a lot going on.

Rich: I mean, let’s say something out loud. Google’s free.

Paul: Yeah.

Rich: Reddit’s free.

Paul: Yeah.

Rich: I mean, Reddit lets you buy, like, badges and stuff.

Paul: Well, everybody has a kind of pay—but, you know, Google, you know, we use Gmail. Like, it’s still not their earner.

Rich: No.

Paul: Like, they don’t—yeah.

Rich: We pay for it, but it’s, yeah.

Paul: It could go away tomorrow.

Rich: Correct. And so these tools, essentially, they situated themselves inside of all information travel on the internet. And then: Plot twist. Was it three years ago?

Paul: Now it seems like forever.

Rich: AI showed up, in the form of ChatGPT.

Paul: Mmm hmm.

Rich: And what’s taken hold is probably one of the fastest-growing disruptive use patterns in technology ever.

Paul: I think it’s faster and more disruptive than social was. Social and mobile were the two kind of before where you’re like, “Wow, that changed the world really quick.”

Rich: People are getting their answers from AI, and they’re going more and more. And it’s not like, “Wow, this is trending.” It’s a billion people.

Paul: And it’s really tricky, because a lot of times they shouldn’t, but it’s a billion people. So you actually have to take—I’ve been thinking more and more about, like, kind of, you need, like, harm-reduction strategies here, because you can’t stop people.

Rich: Yes.

Paul: It’s in the same way that, like, colleges are starting to say, “You gotta use your blue books and write out the essays in hand in the class,” because you can’t be beat this thing.

Rich: Let’s throw out a piece of advice here. Even to this day, it speaks with authority and confidence. It is often wrong.

Paul: [weary sigh]

Rich: It is shocking how often it is wrong. And if you call it out, you’re like, whoa, you’re right. It’ll actually respond back and apologize for being wrong. So here’s—

Paul: We’re going to be saying that, by the way, for the rest of our lives.

Rich: I don’t know about that. I think for a lot of information, I think there’s going to—I think the way Google guides content creators to behave, they’re going to ask for those same… They are in the best position, in my view, to put forward more authoritative content, because they’ve already established a quasi-contract with content creators, and they’ll use that to feed their own play in AI, I think. But I’m diverging here.

Paul: Well, no, I mean, we’ll come back to that in a second actually.

Rich: Yeah.

Paul: So the whole space: SEO. What’s it stand for?

Rich: Search Engine Optimization.

Paul: So if you have been and worked in the web as long as you and I have, it has been kind of background radiation in your life.

Rich: It’s an industry.

Paul: It is. And there are people who are practiced professionals at it.

Rich: Yes.

Paul: And you go to them and sometimes it’s a little dodgy and sometimes it’s pretty legit. And the legit ones are, and I’m thinking of the people we’ve worked with in the past, like, “Here’s how to organize your content. You need a good site map. You need, you know, break content up into these kind of blocks.” It can be a little gross to feel like you’re writing or organizing stuff for the algorithm.

Rich: I mean, you are.

Paul: You are. But it tends to work because, and especially if the focus is like, improve quality and findability, then the robot tends to kind of deliver the stuff and you show up in a good place.

Rich: You want visitors. I mean, that’s just, that’s part of this here. Right? And so, yeah, it’s something that’s been around for a long time. There are professionals, data analysts. There’s all kinds of talent that’s brought to bear data to help with SEO. To clarify for those that don’t know, SEO is you showing up in results that you didn’t pay for.

Paul: Correct.

Rich: You’re coming up in what people typically call organic search results. And that was a big deal. And that’s a whole industry. But the thing is, a huge number of people are now migrating to these other ways of getting answers.

Paul: Which don’t have obvious ad placements. And it’s not clear how to get your content into them. Let’s break it down. So we’re talking about GEO, Generative Engine Optimization, instead of Search Engine Optimization.

Rich: Got it.

Paul: It’s rough. Marketers, when they come up with acronyms, it’s often—like, CAC. It’s bad.

Rich: It’s very unattractive-sounding.

Paul: ARPU. It’s a rough one.

Rich: Yeah.

Paul: Average Revenue Per User.

Rich: Yeah, it’s not great.

Paul: GEO, which means—you know, “geo” is, like, 75 other things, including geography. But no, here we are. Welcome to the rest of your life.

Rich: Yeah.

Paul: So Generative Engine Optimization. Best practices are relatively light on the ground, but I think it’s important to talk about. So I thought we could talk through some of what—I went out and tried to learn a little bit about it.

Rich: Yeah.

Paul: Just before we get into the details, it’s worth noting, this is an era of real complexity between the web and AI. So the best way to see that, and we should talk—let’s talk about Google last, because they have the most sort of skin in the game.

Rich: Yeah.

Paul: But there’s Dia, which is a web browser that has a lot of AI tooling built in.

Rich: Yes.

Paul: That’s by The Browser Company. There’s…Perplexity released one called Comet, their own browser. And these are tools where you go in and it’s, I’ve got access to both of them, and you go into Dia and it’s like, it’s just a prompt. That’s their homepage.

Rich: Yeah.

Paul: And you know, homepages are a funny thing. They’re very, very important. Comet, very similar. They actually really look the same.

Rich: Yeah. It’s an answer machine.

Paul: That’s the new homepage. The new homepage. It’s not a search, it’s an interaction with an LLM.

Rich: So I have two kind of overarching thoughts about this. We’re talking about the web like old dudes talking about the web. But it’s worth saying out loud. Most of the sort of general-purpose uses of AI occur because AI ingested the web. AI isn’t smart, doesn’t have the answers, without the web.

Paul: There is no AI without the web.

Rich: There is no AI without the web.

Paul: Which is actually, I think people need to really bake that in, because what is the web? The web tends to be produced—you know, like, you know who has a lot of web pages? JPMorgan.

Rich: Many thousands.

Paul: JPMorgan—

Rich: I’m sure they’re all really good, too.

Paul: No, no. Merck. Like, these—

Rich: Yeah.

Paul: The web is a place where very large organizations put all their stuff to help everyone else work with them.

Rich: Now it’s worth noting the appetite for more knowledge is so strong for AI companies that they want to take in everything. They want all of the books, they want all of any sort of communication. Like, New York Times has lawsuits out right now because OpenAI—

Paul: They ate the whole Times on the web, according to the Times.

Rich: Right. And Reddit is charging a lot of money. You can actually, you can lease out Reddit’s message boards, all of the comments that are in, across their communities, for money.

Paul: If I’m correct there, Reddit is suing Anthropic for helping itself, but licensing to OpenAI.

Rich: These are the sort of backroom deals that are going to start to happen because—

Paul: I mean, not just start to.There are tons of publishers, tons of news orgs.

Rich: All of it.

Paul: LexisNexis, so on and so forth.

Rich: Right.

Paul: So there’s an infinite desire to squeeze all of human knowledge into LLM-shaped blobs and then give it back to people. But that doesn’t mean that you can put an ad on it or make money off of it in a predictable way.

Rich: Yet.

Paul: Not only that, when it reads the content and absorbs it, or when you have it go do deep research and it pulls the text out and it summarizes it?

Rich: Yeah.

Paul: No ad impressions.

Rich: Today.

Paul: Today.

Rich: Today.

Paul: So, like, the whole ecosystem of Google in the middle and everybody gets a couple pennies here and there—

Rich: Yep.

Paul: And I went and looked at that review of Spider-Man 45—

Rich: Yep.

Paul: And somehow somebody made, like, $200 off that review?

Rich: Yep.

Paul: That’s gone.

Rich: Yeah.

Paul: That is, like, AI has no opinion on that. It’s just like, “I need the text, I don’t really care about the ad.”

Rich: Yeah.

Paul: So it’s a very tricky position, because AI companies want higher-value news sources. The ecosystem of the web was based on kind of, like, we’ll give it away for free, but everybody can have it and we’ll put a lot of ads on it.

Rich: Or subscribe to my, to New York Times or whatever. There’s a couple of models out there. A good way to look at AI in this context is that it’s essentially a really excellent web summarizer. It’s a UX innovation, to a large extent.

Paul: Yeah.

Rich: Because if someone said to me, “You know, what are the five Scorsese movies I should watch?” I could go do that work. I could go read a few pages, go to Rotten Tomatoes, learn, learn a bit about Scorsese.

Paul: You would probably be happier if you did the work.

Rich: You’d probably be happier. You’ll feel more knowledgeable. You’ll stumble on things you didn’t expect to learn.

Paul: I didn’t even know Tom Petty had that album!

Rich: Exactly. And let’s face it, the cat’s out of the bag. Everyone loves the idea of typing into a box, “What are the five Scorsese movies I could watch?” And then it not only gives me a summary, it gives me a little bit of commentary, sprinkles it a bit, has some links in there. It’s really good.

Paul: Let’s be clear, not everyone, because I can just hear, like, there’s, like, five people who are like, “Not me!” But.

Rich: A billion people.

Paul: A lot of people.

Rich: Yes.

Paul: It really is a beautiful combination of, like, human laziness, but it also gives you the sense of having accomplished something.

Rich: Correct.

Paul: So Google, while we’ve been talking, I’ve been having Perplexity’s Comet and The Browser Company’s Dia describe The Aboard Podcast. And they went out to their LLMs and they did a good job.

Rich: Yeah, it’s good.

Paul: They did a good job. Now I’m going to ask Google to do that using its new AI mode.

Rich: Yes, it’s worth mentioning what this is. Google has walked up to the roulette table and bet on every number.

Paul: While you said that sentence. It searched 30 separate things.

Rich: Exactly. Exactly. And the truth is, Google is vast. And Google is very good at what they do. Like, their video generation tool, Veo, is incredible. They just came out with Nano Banana. Do you know about this? And it’s also—

Paul: I do know about Nano Banana, but not everybody knows about Nano Banana.

Rich: Image generation, and it’s extremely good. Google is going to play in this “get summarized answer”—they’re not going to just let this happen. They’re just not going to. Right? In fact, I would argue that they’re going to eventually be the dominant player for these use cases.

Paul: I mean they have the browser and they—

Rich: This, I think if there’s one thing to take away from this podcast, Paul, it is as technologists, we tend to overlook the importance of the default state for most people when they use devices and phones and computers. If the box that you type into is the one that came with your phone when you took it out of the nice packaging, that is the box you will use forever.

Paul: Well, the other thing is that box, the only people who can actually claim the box are really large platform controllers. Like—

Rich: That’s right.

Paul: ChatGPT’s going to make, like—Sam Altman wants to run the world. He’s going to make, like, 50 different plays to see if he can own everything, right? But you got the people we’re talking about here are Apple, Google, Microsoft.

Rich: That’s right. That’s right.

Paul: They’re not going to give up a pixel of that territory without a brutal fight.

Rich: What is the most popular mobile phone platform in the world?

Paul: I’m guessing Android.

Rich: Far and away.

Paul: Yeah.

Rich: Right? You can modify Android. Samsung has opinions about, like, they’ve named their own Barnaby or whatever the hell—

Paul: They want to compare to iOS and iPhone, it approaches free.

Rich: It approaches free. And the truth is, the default builds—and Google did something really smart here. For Chrome, there’s a build called Chromium.

Paul: Rich.

Rich: It’s “open-source.” Put that in quotes.

Paul: Mmm hmm.

Rich: Because they enhance it for themselves. But that browser code base, and then the Android code base, you know, every like year or so they push out a new version of Android?

Paul: Yup.

Rich: Is also produced by Google.

Paul: Do they still name them after candy? I dropped out.

Rich: I don’t know, maybe.

Paul: Yeah. They’re probably out of candy at this point.

Rich: But it’s like, “Wow, Google’s being charitable.” They’re not being charitable. What they’re doing is they’re essentially embedding themselves in the core foundational systems that run these phones. And so you will buy a LG phone which is running Android, because LG just doesn’t have the patience or the resources to write their own operating system. Nobody does. And when you hit that side button, you’re searching Google.

Paul: You know, I don’t think people understand how big—let me, you know how much money all of the real estate in New York City is worth? All five boroughs?

Rich: $11.

Paul: It’s like somewhere between one and two trillion dollars.

Rich: That’s it?

Paul: Yeah.

Rich: Huh.

Paul: Actually, last time I checked, that was maybe 10 years ago. So it was about a trillion then, all-in. So, but that’s what we’re—

Rich: Double it.

Paul: It has not grown hockey stick.

Rich: Yeah. yeah.

Paul: You know, you can’t get a house, but still. Okay. Google is worth more.

Rich: Yeah.

Paul: Okay? So is Apple.

Rich: Yeah, yeah.

Paul: So like, when you think of the entire city, the most valuable thing in New York City is the real estate.

Rich: Yes.

Paul: And our children, who we treasure.

Rich: Hmm.

Paul: But at the same time, each one of these organizations is essentially a giant mega-city with millions of citizens’ worth of company.

Rich: That’s right.

Paul: Like all the land, all the parks, everything.

Rich: Yes.

Paul: Just [dropping a large item sound effect]. That is when you’re thinking about them giving up territory, you’re basically saying, like, “Hey, Google, why won’t you let anyone use some other search?” That’s like saying, “Hey, New York City, why don’t you give up Central Park and everything south?”

Rich: [laughing] Yeah.

Paul: Right? Like, no is the answer.

Rich: No is the answer. Right? And what they understand is that we’re nerds.

Paul: Mmm hmm.

Rich: We like installing new stuff.

Paul: Yeah.

Rich: We’ll install Perplexity on an Android phone.

Paul: Yeah.

Rich: Nobody’s doing that.

Paul: No.

Rich: 90% of people are going to use the really brilliant box—

Paul: It’s tricky, right? Because when you are in your 20s and you love this stuff and you’re learning it?

Rich: Yeah.

Paul: You can’t believe that the world can’t see how exciting and motivating and interesting it is to try something new.

Rich: Yeah.

Paul: And then you go into an organization that has, like, 2,000 people working there.

Rich: Yeah.

Paul: And they’re like, “Please, never, ever do anything new, ever.”

Rich: Yeah.

Paul: Yeah.

Rich: Exactly. And also people in the world, like, it doesn’t have to go into an organization.

Paul: Okay. Somehow, so I asked Google about the AI podcast, or The Aboard Podcast, sorry. And it did pretty good, except it just threw up a picture of Spike Lee’s Malcolm X next to the podcast. When you click through, it’s not there. Like, it’s, I think it’s—oh, you know, it’s an Amazon page. I bet it was like, also on that page. There it is. Okay.

Rich: Oh, weird.

Paul: So still figuring stuff out.

Rich: Yeah.

Paul: But the summaries, they’re all LLM summaries. So basically what’s happening is I no longer—I just searched for The Aboard Podcast. To be fair, we don’t have an ad product. We promote Aboard.

Rich: Yeah.

Paul: And Google made a nice summary of who we are and what we are. Looks pretty accurate.

Rich: Yeah, yeah.

Paul: So do the other two. And there are no real ads on the page except for, like, Apple Podcasts, Amazon podcasts, all the places we’re syndicated.

Rich: Yes.

Paul: So there’s no real money in it yet.

Rich: But—

Paul: Well, I’m assuming they’re going to start slopping ads—

Rich: On Google.

Paul: On Google, yeah.

Rich: See, look, here’s the thing for Google, right? Platforms that capture behavior tend to wait on the money, because if they embed themselves into your lives—I remember when Google started, and it was just so much better. And it didn’t, it wasn’t directory-based and Yahoo was based—it was still years before search ads showed up. They were just capturing our patterns and our behaviors.

Once you have that, it isn’t just about, “Okay, now I’m gonna put a toll booth.” It’s knowing that you have filed away this tool in your long-term memory and it is in your life. Like, it’s over. And then figuring out where you’re gonna grab pennies here and there, that’s not hard. Like, once you have the people, right? And I think Google knows that because they, they did it. They did it once before, right?

Paul: You know, they also, they’ve been slapped across the face. So they try to do their own social network.

Rich: Yes.

Paul: That was a disaster.

Rich: Wave.

Paul: No, Plus.

Rich: Oh, right. Plus.

Paul: That was gonna be their Facebook.

Rich: Yes.

Paul: With circles and stuff.

Rich: Yeah.

Paul: And then AI shows up and everybody’s like, “Where are they? Where are they?”

Rich: I actually have a friend who has since left Google who’s there and he’s like, “I’m not worried about Google. I’ve seen some stuff. They’re being really patient about it and probably don’t want to cannibalize whatever…” You know, they’re still making a ton of money.

Paul: It’s actually—they’re doing fine, don’t worry about it.

Rich: Everybody’s gonna be okay.

Paul: The thing about Google, actually, I would say they launched a lot of their AI products way too soon, because there was so much slop that they let in.

Rich: Yeah.

Paul: They shouldn’t have actually launched when they did because they already—

Rich: What did they even call it? It was Bard for a while. I think they backed off on it.

Paul: But then now they sort of do these automatic summaries and stuff. And half the time it makes it up. It’s not great.

Rich: It’s not great.

Paul: And it’s also at the same time that they’re jamming more ads in.

Rich: Yeah.

Paul: So the core Google experience, and this is a tricky thing, and if you lived—we’re old, so, like, Microsoft in the 90s, they were always putting an icon on your desktop.

Rich: Yeah.

Paul: You’re like, “Hey, I don’t really want that.” And they’re like, “No, no. You really want Novell Netware.”

Rich: Movie—yeah.

Paul: [laughing] Yeah.

Rich: Movie Maker.

Paul: Movie Maker. You want that. You need it.

Rich: Yeah.

Paul: All right, so I think let’s close this out in a very practical way which is, so here we are, okay? We got browsers with AI as the homepage. Google is clearly, like, we know where that’s going to go. It’s going to find ways to put ads in, and it’s going to need to find some way to incentivize content creation so that it can continue to have stuff to feed its LLMs.

Rich: That’s right.

Paul: So we’re hitting that part of the world, and into that comes, now we’re back to GEO, Generative Engine Optimization.

Rich: Yeah.

Paul: And so there’s a couple good articles. We’ll put them in the show notes.

Rich: Yeah.

Paul: But the first one I’m going to summarize and talk about, I’m going to be a human LLM right now.

Rich: Oh, good for you.

Paul: Okay, so this is by Kelsey Libert or [difference pronunciation] Libert in—get ready—Search Engine Land.

Rich: Wow.

Paul: I know, brings you back.

Rich: That’s, like, a 40, 30-year-old website. Has been on GoDaddy for 40 years. That domain name is—

Paul: You can’t name anything Search Engine Land anymore.

Rich: Yeah.

Paul: Like that, it just—so here we are, and it’s a thoughtful article by someone who’s sort of in the industry. And so the numbers are funny because obviously things are nibbling away at Google and people are complaining about the search results, but it still has like 84% of the market.

Rich: It’s going to continue to decline. They know it, too.

Paul: But like DuckDuckGo didn’t eat their lunch. And they’re like, you know, ChatGPT probably won’t either.

Rich: It won’t.

Paul: It’ll nibble. It’ll take, it’ll, it’ll combine, it’ll snap, it’ll snatch something off the plate, like in the high school lunchroom.

Rich: Here’s the thing we often forget, and this is kind of related to my earlier point. Most people don’t change what they do.

Paul: No, they really don’t.

Rich: They don’t. My mom is still typing in Google. Like, she doesn’t—she’s not hankering to get ChatGPT. So most of the world is just doing what they used to do, and it takes a long, long time. I think it was a year ago? AOL internet…

Paul: Dial-up.

Rich: Dial-up! Ended, like, six months ago.

Paul: Yeah.

Rich: There are still tens of millions of dollars a year to be made.

Paul: Yeah.

Rich: On AOL dial-up.

Paul: Yeah, sure.

Rich: Because it worked. And I was like, “Well I don’t need faster. I just want to see pictures of my grandkids,” or whatever it is.

Paul: Imagine that call center. [laughter] You know?

Rich: No, but my point is people embed their behaviors and it takes a really, really long time to change. So people aren’t going to change that easily. That’s just life.

Paul: No. And so there’s a few things you’re going to struggle with when you’re trying to optimize around this stuff. First is when they do their big indexing runs? It’s all at once.

Rich: Yeah.

Paul: So they kind of chomp the whole web. And then they can iterate and update, but it’s not up to the minute. When you put a new webpage up, sometimes Google—

Rich: It’s like a day.

Paul: Google will hit it, like, right away.

Rich: Yeah.

Paul: With AI, you cannot be guaranteed that they will kind of know where your thing is.

Rich: They won’t. Which, by the way, is an argument for Google having an inside track here, because they’re so tuned into indexing topical information and timely information.

Paul: When I did that search, it was so much faster at pulling in stuff from the web than the other—

Rich: It knows everything already.

Paul: They’re figuring out what they’re good at for a while.

Rich: Yeah.

Paul: For a while with AI, they weren’t. LLMs will absorb your content and you’ll ask them, “Hey, what’s Aboard?”

Rich: Yeah.

Paul: And they will take what was in our About Us page and rewrite it in LLM-ese, right?

Rich: Yeah.

Paul: But if I change that tomorrow because I have a really important message?

Rich: Yeah.

Paul: I’m kind of out of luck.

Rich: Yes.

Paul: So I need to really be buttoned up about how I’m communicating and think about the long-term for the content and the strategy that I want to put out into the world.

Rich: Yes.

Paul: Okay? The other thing is that context really matters now, okay?

Rich: Yeah.

Paul: So you can’t just put a page up and cross your fingers or go for PR Newswire or whatever. Things kind of have to fit with other things because that’s how LLMs see the world.

Rich: Well, LLMs translate.

Paul: That’s right.

Rich: They read the whole webpage and then they write a paragraph. So how do you optimize your content so that that paragraph is compelling? I think that’s the challenge out there. There’s this now this intermediary that’s editing your content, effectively.

Paul: So I’m seeing, I’m going over to another article right now. It’s called “GEO Explained in Five Minutes” by Keren Burns, and it’s on storyblock.com. I don’t really know where I am right now, but it’s a good article.

Rich: That’s okay.

Paul: You know what’s wild, though? You know how you implement this? First you write an llms.txt file, which is like robots.txt.

Rich: Oh, interesting.

Paul: And it’s just text. It’s like, “Hey, my restaurant’s open these hours.” And you give it, like, structured markdown.

Rich: Yeah. It probably doesn’t even need to be too structured.

Paul: No. And it just, it absorbs it in its LLM way. And so it’s like, “Okay, this is what they want me to know.”

Rich: Oh, interesting.

Paul: It’s—

Rich: Like robots.

Paul: But flat-text narrative. It’s kind of nice to read.

Rich: Yeah.

Paul: So you can do that. And then, you know where we’re back to is everybody saying, “You know what’s good here? Linked data structures.”

Rich: Yes.

Paul: So for those who don’t understand, there was this idea of, like, a semantic web where everything would be linked and classified by type. So if I was a restaurant, I would say, “I am of type: restaurant. I have a object: menu.” And I would communicate in robot language to the other robots. And that would make a really smart web.

Rich: Yes.

Paul: And that didn’t really happen, but it kind of did. And Google did start to respect that kind of content. So what I’m saying is instead of publishing HTML web pages, you’re publishing data.

Rich: Yes. We’re both been around a while and we thought the web was going to go here. Right? And here’s the best way I can sum it up is that the web is optimized for human consumption, not machine consumption. There’s visual elements that make it appealing for you to look at. It looks like a newspaper, like the front of a newspaper. There’s all the content that you see is for you as a human to consume it.

Secondarily, machines want to consume that content, too. What happened was back then is that Google got really good at parsing human output, the human consumption.

Paul: Well, and getting people to make clean data is, like—

Rich: They gained enough power to be able to tell people to behave a certain way.

Paul: They were the only mechanism that got people to actually produce relatively clean data, because—

Rich: Even if you don’t?

Paul: Yeah.

Rich: Their systems know what to do with it.

Paul: But the reality is that humans making clean, crisp data can only be motivated not by altruism or connection with others, but by the giant platform saying, “Do this please?”

Rich: And then your results will show.

Paul: Yeah, then you’ll get a good menu link.

Rich: Exactly. But what’s happened now, and I think we started the podcast with this point, is that AI is essentially implying that that presentation that you worked on is peeled away. Because I’m just going to read it and put it in a couple of paragraphs.

Paul: Yeah. I want the words. But then if you make nice, individual, crisp data objects and publish them…

Rich: Even better.

Paul: It’s like, “Hey, thanks. Now I know that you’re a restaurant, I know you’re this.”

Rich: That’s right.

Paul: And that gives people the context. So we’re actually back to 20 years ago where you need to be thinking—

Rich: Yes.

Paul: If you’re thinking at scale, according to a lot of the people who are researching this stuff, the more you can do linked data, classic objects, everything gets its own page connected to all the other pages with rich metadata descriptions?

Rich: Yes.

Paul: The more likely you are to be absorbed into the larger context and then come back out through the LLM.

Rich: Yeah, I mean—

Paul: I don’t love it. I don’t love it. I just want to be clear about that. Like, it’s not, it’s not exactly where I wanted to be right now.

Rich: I mean, yeah, I think I get that, I get that. But I mean, spending a lot of time writing a very emotional, appealing About Us page for my restaurant? Does that get parsed and chopped up? Sadly, yes.

Paul: Yeah.

Rich: Because the old way of appealing to others, right? Like, we sell to each other, we pitch each other all the time, right?

Paul: I will put a positive spin on this, which is a web focused on clarity of presentation.

Rich: Mmm hmm.

Paul: I go back to restaurants a lot, because restaurants have the worst web pages of anyone.

Rich: Well, the PDF menu is still a thing.

Paul: It’s still a thing. Remember when they were all Flash?

Rich: Yeah.

Paul: Like, it’s just hard. Restaurants, for some reason, I think it’s just the worst people sell to them. They’re very vulnerable.

Rich: Yes. Yes.

Paul: Right. So it’s just, like, put this on your menu, and then they got, you know, then you got Seamless and ghost kitchens and Grubhub.

Rich: Yeah, yeah.

Paul: But, like, ultimately we’re back to the clearest possible representation of who you are and what you do. You can be as expressive, I think, as you want to be, because it’ll either just absorb it or ignore it.

Rich: Also, for many years down the road, people will still visit your website.

Paul: Yes, that’s right.

Rich: Right?

Paul: And I actually think at some level it’s really in the best interest of the LLMs and Google to start to push people back to the web. They just haven’t figured it out yet.

Rich: I think they will.

Paul: They have to, because—

Rich: I think right now if you see AI results, there’s a little link icon at the end of a sentence?

Paul: Yeah.

Rich: I think eventually the blue link will make its way back because you still want people to go to a website.

Paul: I mean, you have to actually put new food into the buffet. You can’t just have people—the buffet doesn’t constantly refill.

Rich: I think what you’re saying is AI took off because the web was so rich.

Paul: Buffet!

Rich: So we still have to keep replenishing the content on the web.

Paul: You got to come out, you got to put the bacon bits and the peas in.

Rich: My advice is start a magazine.

Paul: Or open a Ponderosa Restaurant.

Rich: A magazine about Ponderosa Restaurant?

Paul: I would actually subscribe to that without a second thought.

Rich: I’m sure you would.

Paul: It would just, you know, like, or one that just, like, really explored Sizzler culture.

Rich: Yeah.

Paul: So on, that note—no, but, like, you know what’s funny is we’re back to, like, really clear quality and clear representation. We’ll put the links to these two articles, I thought they were pretty good, in the show notes.

I don’t think this is as scary as it might sound at first. I think this is going to work out. The web is a second-class citizen in the current context. Like, the broader web of people creating and sharing things that I have mixed feelings about. We’ll talk about that some other time. But here we are. What I have said probably once every month for the last 15 years.

Rich: You’re getting a little—we’re a modern, forward-looking AI podcast and you’re getting a little too sentimental, Paul.

Paul: No, you’re right.

Rich: So I need you to—

Paul: You’re right.

Rich: I need you to man up here.

Paul: For all of my wistful melancholy. [laughter] No, I mean go look at the newsletter. I’m actually, like, I’m starting to see, and so are you. There are patterns for working with this stuff. You know, I just wrote a newsletter article, I hope people check it out, about what guardrails really means when you’re working with them.

Rich: Absolutely necessary.

Paul: And just like how you organize this world. This is not going to be AGI. This is a new kind of technology and it’s going to have to fit in with all the other technologies. And that’s the work of, like, the next decade. And I actually think that’s exciting, good, interesting work.

Rich: And it’s going to take a while.

Paul: And so we’re one of the places that does that. If you check out aboard.com, you can build an app in a couple of minutes. And it’s a good, it’s real software. It deploys to the web. You know, we’re talking to a lot of not-for-profits who are like, “Thank God. I needed a way out. I’m trapped in my systems. And at least this kind of thing is going to give me a way to move forward.”

Rich: A lot of companies, a lot of organizations, hate the software that they have in their hands.

Paul: It’s just there’s so much technical debt in the world. And so there are some really good ways to move stuff along.

Rich: Yeah.

Paul: So I think check out aboard.com we are one answer and there’s lots of answers out there. We’re going to keep exploring.

Rich: We’re the best answer, Paul.

Paul: There you go. That’s our marketing message.

Rich: There we go.

Paul: All right.

Rich: Check it out. Aboard.com.

Paul: All right, friends. And we will see you next week.

Rich: Bye.

Paul: Bye.

[outro music]