# Ep #16 The dead internet, Claude Co-Work Tasks, and a World After SEO

“Ep #16 The dead internet, Claude Co-Work Tasks, and a World After SEO.” AI at Work by AscendAI, 2026-03-16. https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/

Credit the named speaker when quoting a passage, and link to the episode or the passage.

Transcript licensed under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/).

### [Kevin Williams · Passage 1](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-1)

Hey there!

### [Elijah Szasz · Passage 2](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-2)

Woo! How's your weekend, bud?

### [Kevin Williams · Passage 3](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-3)

Another week in AI. You know, I probably need to get a little bit more of a life and not to spend every weekend like going deep into this stuff. I find it is funny. You we have pretty intense jobs and the only time that I really have to play with this stuff and chase my own things are on the weekend. And then what happens is I really grind through the weekend.

### [Elijah Szasz · Passage 4](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-4)

You did it again, huh?

### [Kevin Williams · Passage 5](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-5)

And I totally crash on Mondays. So I go to bed at like, I don't know, eight o'clock on Mondays because I'm so tired from my weekend.

### [Elijah Szasz · Passage 6](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-6)

Yeah. Yeah, I totally feel that. I intentionally, well, I'm not going to pretend like I did this my own volition. I promised my wife I would just stay off the computer this weekend. And I really had to spend some time with the kids and the family, ⁓ just get some sleep. And I tell you, I slept so well. I think there is something about the beginnings of World War III, autonomous killing machines, and mass surveillance that just I skipped the light sleep and I just go right into REM sleep or something. I don't know what it is. It was amazing. It was amazing. And ⁓ you know what, Kevin? When we started this, we're like, we're not going to talk about just all of the churning news around AI because it's always such a frenzy that it's kind of silly to try to keep up with and everybody's funding footprints. And we also are not going to get like into all the different political sides now. And we're in just such weird times. I think we just have to at least make mention of what happened. We don't have to do it right now, but there's just so much crazy stuff going on. you know how we were talking about Matt Schumer's essay last week or two weeks ago, or I don't even remember when because it's such a blur. And the thing that you and I had connected over was, wow, we both had that kind of feeling that this feels like the early days of COVID where there's all of this data. And you can kind of project out what the future holds. And it's really hard to ignore it. And then people look at you like you're crazy for talking about it, right? And how that just felt so much like AI right now. The other thing I realized that happened with those early days of COVID is that there's this epidemic where people's lives are in danger and people's livelihoods are in danger. Entire companies are in danger. Yet the entire thing just went from just green to red line in terms of being politicized, right? And I feel like that's starting to happen too, right? Like it's really hard to start to have these conversations without it then being a political conversation as well. And I don't know, sometimes I feel like I'm taking crazy pills where has it just always been like this and now I'm much more aware of it or does it seem to be heating up? And that's just the world we're in now.

### [Kevin Williams · Passage 7](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-7)

So a couple of different reactions. ⁓ One, you ain't seen nothing yet. As we're heading into whatever side of the spectrum you're on, this is going to be a really special year from an election perspective. you know that I spend a lot of time quietly briefing lawmakers and policymakers, et cetera, on the practical boots on the ground aspects of this. And these are really smart, dedicated people like forget politics, red, blue, whatever. These are people who care. They totally care. And they totally don't get it yet. ⁓ as much as I'm trying and as much as other people like us are trying, they have, ⁓ sort of, hope that it's a fad like, ⁓ you know, I've seen some, we've tried it and it didn't work. it's like, still.

### [Elijah Szasz · Passage 8](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-8)

Really man, still? That you're still getting that? Okay.

### [Kevin Williams · Passage 9](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-9)

I'm still getting that and I'm still getting that from very senior people. And there is this lensing effect, I think, and this isn't just politicians. This is anybody who's senior in an organization. And the bigger the organization, the worse it is that these people are just so crazy busy with running their public company or whatever it is that they're heavily reliant on other information sources that are coming from within their own businesses, right? And in a lot of cases, sure, that I can't give you a specific example, but I

### [Elijah Szasz · Passage 10](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-10)

Can you give a quick example of that? Not specific, but just kind of, ⁓ yeah, like a generic contextual thing.

### [Kevin Williams · Passage 11](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-11)

So, okay, public company CEOs, I do work with a few as sort of a personal advisor, which is an interesting role to have for sure. And what I'm finding is, it's very, very intense to have those jobs. You're beholden to public shareholders, you're beholden to the media, you've got to be very careful what you say in public, et cetera. it's just a full on type job. And the nature of those bigger jobs is that they're also surrounded by a lot of other people, okay? And those people, I'm not for a second gonna say that they're ill meaning, but they're also super busy. And in a lot of ways they have their own lenses that the reality of the boots on the ground, like successful project type stuff that you and I are seeing, isn't really translating up to them. So they're getting a bit of a negative impression of the way this works. And also it just upsets the order that even a very tech forward CTO, this is a lot to deal with as far as data structures, as far as security, as far as access, et cetera, et cetera. And it's sort of human nature to not want to change. And if you end up with an apparatus below you in leadership who doesn't totally buy into it or isn't getting it, then the information that's getting to the top suggests that this isn't really as much as it's cracked out to be, which creates a governance problem that doesn't create the right resources or the right agenda in the organization such that floats down. This is precisely why I'm always advocating this idea of a manifesto. And I know I've talked about it. Yeah.

### [Elijah Szasz · Passage 12](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-12)

⁓ man, took the words out of my mouth and I have some stuff for you on this too. Yeah.

### [Kevin Williams · Passage 13](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-13)

that a manifesto coming from the very top of the organization, not the CIO, not the CTO, but the chairman, the CEO, the president, the founder, whoever it is, who is making a public statement, whether that's carefully worded from a PR perspective or not, but as far as organizational alignment, it's basically saying, we believe in this, this is going to change the way that we operate, maybe in some big ways, maybe in some small ways. And as a team, we all need to figure this out. And that creates the alignment with the rest of the leadership team, who then is going to go deeper into the organization. And you're not having that weird conversation about facts. Like, people will always debate facts, but there are some facts here that I think are pretty inescapable. And if you can get beyond the surface level debate of is this a thing or is this not a thing, that's step one. And a ton of organizations have not gotten over that first part.

### [Elijah Szasz · Passage 14](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-14)

Yeah. Yeah. Well said. Well said. And it's the little story that happened over the weekend when I was not working yet, still talking about AI as one does came up and we can circle back to the whole anthropic and Pentagon debacle and how open AI spooked in and what's good. mean, listen, it's all kind of still happening in real time, even though some stuff has decidedly happened and So it's a weird time in US history for many things that included. ⁓ But let's come back to it because I mean, honestly, we could probably just spend three hours talking about this and ⁓ I'm gonna see you for the first time in person for, I don't know, like months tomorrow. So maybe that's what we'll do is just talk about this for three hours tomorrow when I meet up with you for coffee. But so I wanted to get your take on this because I thought it was interesting. I had a friend come over who works for a bank that I won't name. It's not the. biggest in the food chain, but top 40 bank that I think you and I will probably be talking to very soon. And there is this really interesting take on AI adoption from a leadership perspective. A lot of the tripping points that we talk about a lot, one that you just brought up, like no real AI manifesto coming from leadership, even though the entire organization is talking about it, right? No AI ambassador council. Some co-pilot licenses haphazardly just thrown out to the organization with no training or workflows or real instruction. She knows there's tons of shadow use of AI, which we've both talked about before. Shadow use being the use of AI in the workplace that isn't sanctioned by leadership. And employees are doing it because, I don't know, it's smart probably for their role to be using these tools because they can have much more output than if they don't use these tools. But I thought this one was really interesting because I've been hearing it more and more and have even experienced it to a degree in my own business. So I'll put forth a single use case that you can kind of me your feedback on. Customer success call centers, right? For most big companies today, and I'm friends with some of the people who run these services overseas, all of this stuff is essentially routed outside of the US. These massive call centers, labor is super, super cheap. And this is all ripe for AI disruption, right? It's already been proven time and time again, and lot of companies already adopted this, that AI voice agents and generative AI can very often do a better job than somebody who's unskilled and doesn't have all the knowledge or easy places to access it ⁓ to provide really good customer success, right? But in this case, she made this argument that at least from the management's perspective, The costs are so low for this service as opposed to doing it in the US, right? But hear me out, like the costs are so low that to even get the workflows built to get through that initial pain of adoption and start to embrace this stuff and implement it, that leadership can't even get their heads around taking on this initiative. And it's just a really strange thing to think about because obviously if you project out the economics on a long enough timeline, of course, this is just going to be pennies on the dollar, When it all works. But I've heard this before. I've heard the same thing with data cleanup, right? we want to build out all of these workflows and connect all these dots or have this knowledge graph that we've talked about before. But the... effort and the cost of initially doing that. And I think this just ties so perfectly back to what you just said about how a lot of people in leadership still don't believe what is about to happen, what I think is undeniably about to happen and still think that, maybe this is a little fat or bubble that's going to pop and then things will just go back to normal. But I'm just really curious to get your take on this because it isn't the first time I've heard this story.

### [Kevin Williams · Passage 15](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-15)

Okay, so ⁓ I wrote an essay about this a couple weeks ago ⁓ that I have some theories about disruption in SaaS in particular. And this is sort of me wrestling with as like a retail investor, like should have invested in video, didn't, but you whatever. But now looking at like my own holdings and trying to figure out like where my exposure sits and.

### [Elijah Szasz · Passage 16](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-16)

Right. Yeah.

### [Kevin Williams · Passage 17](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-17)

trying to break that down into like who's defensible and who isn't defensible and why. And before I get to your key point, the first two are how close is the offering, particularly in SaaS, the software as a service to enterprise, because enterprise moves really, really slow. It's really, really cautious. Once you're embedded, everybody knows that you're probably going to stick around for a long time, even if your platform is terrible. Then the next is how close are you to regulation? How close are you to doing law, to doing finance, to HR, like things where there's real regulation. That creates some stickiness as well ⁓ as far as slow moving. And then the last, think, what I think I'm gonna use to address your question, which is, is the function you offer ancillary or core to the business? And this is different on a spectrum of businesses. So if you are a giant bank, and you offer customer service, then customer service, although maybe expensive and maybe painful is really far away from your core activity. mean, certain banks, certain activities, like, you know, there's a lot of nuance here, but, but it's not your actual business, right? Okay. But now you're a debt collection business and your entire business is

### [Elijah Szasz · Passage 18](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-18)

It's a must have, but it's not your actual business, it's, but you have to, but you have to have it. It is. Yeah.

### [Kevin Williams · Passage 19](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-19)

predicated on being able to reach as many people as fast as you as you possibly can. Now, when you're looking at your PNL or you're looking at your budget, that call center is an enormous part of what you do. So two different businesses essentially running the same team in the Philippines. One is looking at it as, ⁓ you know, this isn't going to really change the way that we do things and people aren't going to love it. So. How about we focus on, ⁓ I don't know, like crypto tokenization, stable coins, like something that's like this big deal that's gonna change the nature of our business. And whoever runs that P &amp;L is gonna be looking at it like, ⁓ maybe we can save money, but like the ROI isn't like shut up and take my money. It's gonna take a little while because of what you're talking about. But again, debt service business, if it's 30%, 50 % of your expenses, you're gonna be all over that thing and deconstructing it. ⁓ From the investment lens, ⁓ I wrote the essay basically about Adobe and like the contention that with all of these new, cool, creative tools that are coming out everywhere and how good CapCut's getting and how good Canva's getting like, ⁓ my gosh, this is going to eat Adobe's lunch. And then it's like, Hey, you're this big bank and you've got a hundred people with Adobe's Adobe licenses. Yeah, that's pretty expensive, but it's $10,000 a month expensive versus everything else you have going on.

### [Elijah Szasz · Passage 20](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-20)

Right.

### [Kevin Williams · Passage 21](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-21)

but you're a creative agency and you have 100 people with these licenses and it's ⁓ it's a.

### [Elijah Szasz · Passage 22](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-22)

The most expensive and the most expensive ones, right? Like top tier licenses. Yeah.

### [Kevin Williams · Passage 23](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-23)

Yeah. And it's the top tier licenses. It's really expensive. You're going to be all over this. What does that mean for Adobe? I mean, it's not great for Adobe long-term, but it doesn't mean Adobe is going to crater right away. There's a lot of stickiness and a lot of value in not, not breaking things that already work. Like we don't need to break everything. And I think we both see this that AI has given organizations carte blanche to go in and start turning over rocks and they're, they turn over 10 rocks and lo and behold,

### [Elijah Szasz · Passage 24](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-24)

Right.

### [Kevin Williams · Passage 25](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-25)

Seven of those rocks aren't AI problems. They're just problems that have been hiding there in the shadows somewhere. And it's a process problem. It's a human problem, whatever it is, not an AI problem. Or, yeah. Yeah.

### [Elijah Szasz · Passage 26](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-26)

or their P and L problems. They're just big line items on the P and L. And I think the example you brought up, and we talked a lot about this on the last episode where you spent the weekend vibe coding what could very well be a formidable SaaS product out on the market. ⁓ Say you're one of these organizations that uses a very expensive CRM like Salesforce, right? ⁓ Very few of the Salesforce implementations are actually using all of the Salesforce features that are being paid for. Most of it is really just bespoke and in fact, it's so complicated to implement that the whole cottage industry spun up of other companies that just will implement the CRM for your organization and that's expensive as well, right? So we're reaching this very strange time where leadership is starting to think huh I could just bring in a couple developers and by developers I mean vibe coders to create a bespoke version of what we currently have implemented. And all I'm going to be doing is paying their salaries. There won't be any third party implementation costs. We're going to get exactly what we want. And this thing that is now a $200,000 line item on the P &amp;L every single month is just going to go away and be replaced by a couple hundred thousand dollars in salaries. And that's a really interesting thing to think about that's probably right around the corner as far as those types of decisions being made. Have you had any of those conversations?

### [Kevin Williams · Passage 27](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-27)

It's funny, I was on a call with a not-for-profit yesterday and they were asking, okay, we're off, we're going, we're starting to understand this, thank you very much, but how do we develop this capability? And I sort of had a pause. I'm looking at this room full of not-for-profit leaders and they're earnest and they want to stretch their resources as far as they can. And what I told them is you need to find somebody in the organization who has the curiosity and the interest. and you need to carve out part of their role to be like not the AI czar so much, but like the functional person who can help a team who now understands the strengths and limitations of the technology, turn that into a project that makes sense. And because if you leave it in the hands of these 10, 15 well-meaning people, they're probably not gonna build that much. They're not gonna be able to individually navigate the data connections and security issues. ⁓ This organization works with children. You've gotta be careful about this. So how do you do that? Well, you have a point person who, be they a developer or be they a vibe coder or be them just a more curious than your average bear person, like that they're the person who does it. and then they can get the right permissions and have the right data access and the right curiosity and make things happen. But at least everybody's using the same vocabulary. When you have shadow AI going on and Joe in accounting is doing one thing with Gemini and I'm doing something else in marketing and whatever, it's terrifying because people are tapping into things left, and center. ⁓

### [Elijah Szasz · Passage 28](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-28)

Yeah. Yeah.

### [Kevin Williams · Passage 29](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-29)

But you formalize that based on the size of the organization. If you're a 500 person organization, it starts with somebody who's actually of leadership rank who owns it. And then they start building out their team. And maybe it's just part of those people's role, or maybe it becomes their whole role once they have the right skills.

### [Elijah Szasz · Passage 30](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-30)

Yeah, yeah, it's super interesting. And it's moving really, really quickly as well. ⁓ Kind of on the topic of you VibeCoding a SaaS product over a weekend that we talked about last time, I think pretty much for the entire time we talked about you VibeCoding that product. We had also talked about my follies with OpenClaw and you had tried it as well. That's also made quite a bit of news.

### [Kevin Williams · Passage 31](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-31)

You

### [Elijah Szasz · Passage 32](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-32)

especially since the founder of that open source project is now working for OpenAI, rumored to have gone over there for a billion dollar salary, that's billion with a B, which is quite the signing bonus. I'll take that one all day long. Kevin and I got into this little bit of a debate. We both jumped on right when the news broke of all the amazing things that this agent in a box could do and we both had some trouble getting it set up. And then I decided to go deeper. Kevin decided to watch me flail, which I hate admitting you're right, man, but I should have just waited. I went through so much hell, buddy. I bought a Mac Mini. And then because the news was moving so fast and this thing was getting updated so quickly,

### [Kevin Williams · Passage 33](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-33)

Hahaha!

### [Elijah Szasz · Passage 34](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-34)

It's difficult to even find a YouTube video that's giving you the current information that was posted two days ago because everything has changed. And so I decided to set up a custom gem with all of the information to act as my tutor to guide me through this. And it did a pretty good job. And it also led me astray with a bunch of stuff. It told me to first run this thing headless, meaning that you don't have a keyboard or a mouse or a monitor set up on this machine. but you're tunneling through a screen share from a computer that you're already using, which in this case was my Mac Studio. And to establish that first and then set it up and create a unique Apple ID and just a lot of just ⁓ IT administrative hoops to jump through. And I just had gateways crashing and ports getting blocked and this just went on for... Enough time that I'm embarrassed to even say what the total time was. Like this just went on forever. And then I started asking questions of Gemini. Like, do you think trying to do this through the screen share is problematic and that's what might be causing this? ⁓ you're absolutely right. Eli, should be working directly on the machine. I'm like, great. Thank you very much. And I eventually did. I worked directly on the machine. I got this thing up and running. I had one last bug. just couldn't get around and I then asked Gemini again. Do you think just this might sound silly crazy idea, but hear me out What if I just restarted the machine do you think it might fix some of these problems? I think that's a great idea. I'm not kidding. I restarted the machine like the old just Everything worked flawlessly it it worked perfectly right that said

### [Kevin Williams · Passage 35](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-35)

That's a squeeze edit.

### [Elijah Szasz · Passage 36](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-36)

It's still everything that you want to do from there on out, every skill you want to implement, every workflow you want to try. It's still a pretty heavy lift. And then I got a message from you saying, hey, have you checked out these tasks on Claude Cowork? Kevin? Did you end up playing with any of the tasks? Just out of curiosity, because I know you're already deep in Claude Code and building all sorts of random lead magnets and stuff through that. So I didn't know if you had the bandwidth to dive into tasks.

### [Kevin Williams · Passage 37](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-37)

⁓ I haven't, I haven't to the degree that I know you have. ⁓ so these, but, but the idea is they're scheduleable. So if you want to go out and say, you know, look at a site or I don't know, check sports scores or whatever you want, it's going to do something that simple to sort of on its own at like 7 30 AM every day. then because Claude now has connectors, which we've talked about before, which part is part of what brought on the SASpocalypse. is that it's trivial, and I mean trivial, to then post it to Slack or move it to ClickUp or do something in Drive or whatever it is. So these are little, very easy to build agents, agent structures, and they can be super simple or they can be really complex. And I've seen some people doing some wild stuff, including you.

### [Elijah Szasz · Passage 38](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-38)

That's right. Yeah. Yeah, yeah. So in GeekSpeak, these schedules are called a cron job, right? And this was one of the things that everybody was talking about with OpenClaw, that there was this 3 AM protocol that went kind of viral, allegedly some of the agents decided on their own to start doing work on their humans behalf at 3 AM. Because why wait for instructions? We already know the outcomes that this user wants. Well, to start building things at 3 AM. The reality is, that is a cron job that you set. You could say, hey, based on all of the conversations that we had today, surprise me in the morning, start a cron job at 3 AM. I'm going to check in at 7, and I'm going to see what you've done. So yeah, that is now possible inside of these co-work tasks. Other things that are possible that were also happening with OpenClaw is that it's really, really simple. to set up something like WhatsApp or Telegram, Telegram being the easiest, so that you could have conversations with these little agents just right on your phone in Telegram and say, you know, I have a random idea. Hey, could you please do this, this, this, and this? It calls the tools that you set up and then automagically it happens. Or you could say, hey, every morning give me this report with all of the AI news and tell me about my Google Calendar and anything in my inbox that needs attention. And the guardrails appear to be a little bit more substantial and better in place than the Open Cloud project. I personally feel a lot safer, even though you are still, as you these up, saying, yes, I understand that this is risky. Yes, I know that generative AI does spooky stuff, and we don't even know why. So all of that stuff is still happening. But I'll say that one of the tools that is maybe going to be a game changer for me is that there is a Chrome extension, specifically published by Anthropic for Claude that you can install into your browser. And then you can say, hey, do this task on the internet for me, right? ⁓ No, need a specific tool call for WordPress or Squarespace or for Riverside that we record this podcast on. It's just. any website ever. And then you can have a set of instructions with tasks that it performs at a certain time, at a certain sequence, and references a bunch of files for you, and can basically take the place of you, the human, clicking all those buttons through the web to get a task done, which I don't know about you, Kevin, but it takes a shocking amount of my day to click buttons on the web. Like I'm clicking, I'm clicking a lot of buttons on the web every day. So when I think about things that can give me back my time, This kind of blew my mind. And I saw early signs of this. We both talked about agentic browsers, for example. So a lot of the Frontier Labs came out with their own browsers. Kevin and I jumped on right away. were like, oh my gosh, this is the greatest thing since sliced bread. And then 20 minutes later, we messaged each other. like, yeah, this is, none of my plugins are working. Oh, this just broke my whole workflow. It keeps crashing. Never mind, right? But this is using the native browser with all of the stuff that you use every day just as a default. it's a very, very different experience. And while it's not quite there as far as being flawless, like I was blown away. I spent a lot of time playing with this. And full disclosure, the thing that I was trying to do, which maybe I'll talk about on a future episode, but the thing that I was trying to do with OpenClaw that I dedicated, days of my precious life too, and never really got to work because it just kept crashing and faulting out. I did in this format in roughly 45 minutes. Like roughly 45 minutes, it was all working. Like working really well. anyways, ⁓ yeah, yeah, when we kind of talk about AI at work, you know, one of the things we always talk about are what are the tasks that you and your team really despise doing, which just feels like bullshit work, pressing buttons, moving things around on different little cells on a spreadsheet that you could possibly use AI to replace you with in that particular use case that you can spend your time doing better things. For me, really opened my eyes as far as, it's not just around the corner. The future is kind of here, and I just expect it to get smoother and faster from here on out.

### [Kevin Williams · Passage 39](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-39)

Thank

### [Elijah Szasz · Passage 40](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-40)

But I can't wait to finish up this little show so I can just jump right back in, because I have so many ideas of just different things that I want to experiment with now.

### [Kevin Williams · Passage 41](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-41)

So unfortunately though, there's going to be this window of opportunity where this stuff is going to work. And what you're leading to are two different concepts. One is, they call it the dead internet when it's basically just agents, like interacting with each other and creating like slop based on each other.

### [Elijah Szasz · Passage 42](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-42)

Which to be fair, we're kind of seeing the beginnings of because you and I, talk to marketing agencies all the time and there's now entire farms of AI personas and avatars that have their own Instagram accounts with over a million followers. They are literally AI influencers. They are not people and they're interacting with users and other AI agents.

### [Kevin Williams · Passage 43](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-43)

We're seeing the beginnings of this already. It's. It's wild. And, just, just, ⁓ this just popped into my head. So Metta, Metta just filed a patent for, ⁓ the ability to have pervasive AI experiences with literally dead Facebook users, ⁓ such that you can continue to talk to whoever in their Facebook channel and they react. And because there are Facebook has now been around long enough and the demographic is old enough that there are basically

### [Elijah Szasz · Passage 44](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-44)

We have more dead users than alive users. Not creepy at all. Not creepy at all, Cha-ching!

### [Kevin Williams · Passage 45](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-45)

They have more dead users than live users. And if you're, if you're Zuckerberg, you're like, Oh, that's an opportunity. So, uh, you know, like, let's talk with grandma and, but that's good. They're going to create hi. They were really going to do this there. I mean, maybe who knows? I wouldn't put anything by him.

### [Elijah Szasz · Passage 46](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-46)

Yeah.

### [Kevin Williams · Passage 47](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-47)

⁓ that it's all about time, right? And if you can create like engaging activities in a way that's new and different, that will get the eyes, the eyes will get the ads, the magic money machine keeps running, right? So dead internet is one thing. So, you know, from an AI at work perspective, you have to be injecting yourself and your real self into that content. And like the way that that's what we're doing right here in part is we talk about this a lot. But this is the summation of our week of AI thought. And then now this is like super interesting fodder for us to be able to talk about in essays and other things that didn't come from the AI.

### [Elijah Szasz · Passage 48](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-48)

Kevin, we have a whole AI-generated document of talking points that we just ignore. We don't even use them. But they're there. They could be there. We could use them.

### [Kevin Williams · Passage 49](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-49)

We always ignore, we always ignore, but which is perfect though. We should ignore it because this particular take doesn't exist anywhere else on the internet. Right? So that's not dead internet. That's fresh internet that we can inject using our various processes. So if your marketing team is out there just creating a whole bunch of swap blogs, you're not doing it right. You're just playing into the dead internet and you need to be figuring out how you put like contemporary information and founders hate this, but. founder engagement and founder interviews and things like that work because it's like that hot take from the CEO is not available anywhere in the LLM, right? ⁓ So that's the one concept is the dead internet. The other is the zero click internet and the zero click feature. And as you know, having a digital marketing background, this one should give people shutters that ⁓ it's the idea that it's not at all about attracting me, the user into the environment. It's instead about attracting the agent into the environment. And the agent doesn't care about your pretty pictures on the site. In fact, the agent can't even really make sense of what's on your product page or your information page because it hasn't really been trained on it. Not in the way that people think it has. It doesn't just like suck in the page and have every word. It's sort of a gist of what it is. And the gist is not enough. So you have to lean into giving the agents the data layer that they need in order that they can grapple with it. And two sides to this coin. One is just like the tactical or the strategic. And we're working with clients trying to figure this out and what that data layer needs to look like. And then there's just the performance part. So you spend all of this time on this beautiful website. And if you are in a transactional space, nobody's going to care what is on your website because they want to get to their thing as rapidly as they can. like

### [Elijah Szasz · Passage 50](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-50)

And to be fair, even if you're not in a transactional, well, mean, listen, if you're in business, you're in some transactional space, right? But what you're talking about is e-commerce specifically, right? Like I am going to go no or yes. OK. OK.

### [Kevin Williams · Passage 51](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-51)

No. Both. It's not, don't want anybody to walk away with the fact that it's just e-commerce because this is important. it is, yes, e-commerce is super obvious because if you're selling a bunch of and all the pens are kind of similar, like, eeh, nobody really cares too much. They just want to buy their pens. If your child is selling a travel experience or something that's like a richer, like part of the purchase is actually part of the experience, you're in a better place.

### [Elijah Szasz · Passage 52](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-52)

That's what I was going to say.

### [Kevin Williams · Passage 53](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-53)

but we transact so much on the web. Think about like a dentist website. you know, you want to find the best dentist as fast as you can and get on their schedule as fast as they can. You having to mill through the pretty pictures of happy, you know, young families on the dentist site doesn't add a whole lot of value if you trust that the LLM is going to yield the right solution. And then the same transactional element is the handshake between what's happening in the LLM the agent layer and how that leads to booking ⁓ an appointment based on their backend booking system such that you never visit the site.

### [Elijah Szasz · Passage 54](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-54)

That's right. That's right. It all, it all happens after you just go into Gemini and type, I am looking for a Dennis in this area, yada, yada data comes up. There's a booking button right there. And that's it. That's, that's the way, that's the way it happens. Right. And it's really strange to think about. And it just, we talked a bit, we actually talked a lot. think about this maybe last week. I don't know. Memory's kind of blurry, but That whole zero-click world, really is starting to remind me of, with OpenAI's announcement that they're gonna start putting ads into the results of different people's prompts. A whole other can of worms here. But in the early days of search, if I wanted to find something, if I wanted to find goods or services before we had this whole secondary search universe just in the Amazon search bar, You go to Google and you say, hey, I'm looking for this thing. A bunch of blue links come up. And then over time, everybody started noticing that all of those blue links that came up in the first page were there because brands paid for them to be there. That's the way it went. So it went from this whole kind of organic discovery experience to who's going to throw the most money at this problem from a brand perspective to be in the top search results. And right now we're like in those early days of Google, doesn't it feel like that with the LLMs where you can ask these questions and have these conversations, put in these different prompts, discover things, recommendations come up and you're like, wow, it's just kind of magic. The model must know it did all this research and it came back and it gave me this stuff. How long until that stuff that it gives you is just a result of who paid for it to be put in front of you, right?

### [Kevin Williams · Passage 55](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-55)

So there's a bunch there ⁓ about whether or not ads are really going to be a thing from a trust perspective. I think they will to a degree, but this is a real problem for the platforms themselves, given the cynicism of most people and how that's going to work. And the easy switchability, like GPT starts doing it and you feel like it's disingenuous, you're going to jump ship. And the fact that there's plenty of money to be made outside of ads. I know that's like blasphemy that, Google makes $200 billion a year basically just on ads. Google makes, has an opportunity to make a lot more than that by connecting the dots behind the scenes with all this other stuff. Google does not have to be an ad company anymore. And I think that the assumption that that is their motivation may be incorrect. The other piece, if you scroll back the clock, I now it's gonna make us sound old, but all the way back to like 2012.

### [Elijah Szasz · Passage 56](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-56)

Yep.

### [Kevin Williams · Passage 57](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-57)

when Google did kick in with which oldsters will remember the Google knowledge graph days. And the other thing that I think is happening and I think that our listeners should file away is that there's this sort of wait and see attitude about like, okay, there are all these standards. This is really complicated. I just need to chill and wait. And the internet's going to tell me like which direction I need to go with this. Well, what happened back in 2012 when Google clicked in with its knowledge graph is that all of a sudden the people who adhered to these new standards got an immediate advantage. And Google told everybody that this was gonna be the case, but they got a durable advantage because once you got that flywheel turning, that the results were yielding more results, were yielding more visitors, were yielding more interactions, you were rewarding the Google monster. And the Google monster said, this is great. The same thing is happening in the LLMs.

### [Elijah Szasz · Passage 58](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-58)

Red. 100%.

### [Kevin Williams · Passage 59](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-59)

Everyone is giving the LLMs too much credit for differentiating between different types of valuable traffic. And the reality is they don't care. Nobody's really paying attention to that on that granular level. So my supposition is that brands, be it an e-commerce company, be it a dentist, whoever, that gets the right reinforcement mechanisms in early such the LLMs can gravitate to it, that's going to send a signal in the next training run. So we do a training run every six months or so. And in the next training run, the winners of the previous training run are going to be bigger winners. And it's going to be hard to break into that as a brand, particularly brand. That's not like a real brand, like a global brand. So unfortunately the, the, the, the, what we're telling people to do is to place a lot of bets on the different types of schemas that are out there such that.

### [Elijah Szasz · Passage 60](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-60)

Yeah.

### [Kevin Williams · Passage 61](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-61)

You know, some of them will work, some of them won't, but if you're not there at the party, you may never be at the party.

### [Elijah Szasz · Passage 62](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-62)

Yeah, it's funny. We have some product, some sites deployed on ⁓ Squarespace, the no-code platform. Now it's just all basically a framework for custom coding that Cloud Code is putting together. But I didn't notice in their admin panel that a new little feature had popped up saying, ⁓ optimize your site for answer engine optimization, right? And it kind of takes you through this little wizard of all these little things you can do to try to get your pages surfaced in these different LLMs. And it's interesting because a lot of it right now is speculation. I mean, even if you ask some of the actual machine learning researchers inside of these big labs, well, how exactly is this making its way out to the answer of somebody's question that they put into a prompt in ChatGPT?

### [Kevin Williams · Passage 63](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-63)

Mm-hmm.

### [Elijah Szasz · Passage 64](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-64)

They'll say, yeah, we don't really know. We have some ideas. We have some kind vague guesses. And one of the things, and you can weigh in on this, Kevin, that has been speculated or clearly defined now, you tell me, is that the way that these models want to return information to different users is in this question and answer format. And you'll see this growing trend. all over the web on people's different pages of FAQs, right? Every page is now amended with these FAQs. ⁓ They might even just take all of that page's content and then put it in a question and answer format because allegedly, this is what all of the training monsters are out there sniffing around for is like, here's the question, here's the answer, here's the question, here's the answer. They connect those dots and then regurgitate all that to the user who said, where can I find the best water bottle that keeps my drink coldest, the longest, and is super rugged. And if you've asked and answered enough of those questions about that specific thing somebody's looking for, you have a better chance of it surfacing when somebody poses that question in a model. Have you heard something similar?

### [Kevin Williams · Passage 65](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-65)

Sure. I, I'm going to say something sort of, mean, I don't know. Like I, I'm obviously surrounded by a lot of different marketers. ⁓ I actually differentiate between AEO answer engine optimization, which is basically what you're talking about there and on a, both a platform basis. So you've got Google's AI answers. That's actually yielding more LLM results than anything out there because people

### [Elijah Szasz · Passage 66](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-66)

Do you think it's real? Are you talking about just a little Gemini results that pop up when you now type anything into Google? Okay.

### [Kevin Williams · Passage 67](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-67)

Yep. Yep. That's AI answers and those snap to things like that. then the FAQs and the accordions, I mean, this is all helpful, but what I don't necessarily buy is that it can be random, that it needs to be more structured in a way that, it's pulling from set data. All right. So, and then the other part of it is, is GEO and GEO, I, this is all you'll talk to one person who will limp link. lump all of these things under AEO and another person would link all of them under GEO.

### [Elijah Szasz · Passage 68](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-68)

This has changed names a million times already, everybody, by the way. Very fluid,

### [Kevin Williams · Passage 69](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-69)

Yeah, has. it still will. It's like SEO, search engine optimization, that eventually was the term that everybody landed on. I still like to differentiate them because they're different tactically and they're different strategically. That GEO has a lot more to do not just with, it's closer to SEO as far as the stuff that's going on on your site, but it has a lot to do with the influence of the broader internet. So Reddit and reviews and things like that. ⁓ The reason why this is important as far as structural data is because the GEO layer is looking to fill categories that were never a thing before. It's not just like black pen with blue ink or, you know, that sort of stuff is being carried in your product feeds. ⁓ It is about sentiment and why people are buying this. And you actually are starting to see fields like occasion, sentiment and relationship. So like that way, you know, it's a, work with a, with a really cool gifting company that sells sympathy gifts. And if you're in GPT and you're like, ⁓ my mom fell and she broke her hip. What would be a really thoughtful gift? And so, so it's matching up in the data of occasion.

### [Elijah Szasz · Passage 70](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-70)

What do I do? What do give her?

### [Kevin Williams · Passage 71](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-71)

The occasion is sympathy, like something bad happened. The sentiment is obviously like love and affection, familial relationship, and the relationship is familial as well. And if that's been pulled from all of these other sources across a big catalog of activities, then it gives traction points for the LLM to latch onto it. The fallacy, I think, and you know, not,

### [Elijah Szasz · Passage 72](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-72)

Mm-hmm.

### [Kevin Williams · Passage 73](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-73)

I'm not the data scientist, I'm not sitting behind the scenes of the FAQs, is you're assuming that the training run is looking for all of that information and is actually correlating all of that information and pulling it into its hive mind in a way that it's connecting, I think it's connecting on a specific skew level. not trying to get wonky because a skew can be a dentist appointment, like an event. Let's just call it an event, not a skew.

### [Elijah Szasz · Passage 74](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-74)

Sure, sure, any transaction.

### [Kevin Williams · Passage 75](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-75)

And if you have a red pen versus a blue pen, the accordion stuff is going to be very, similar. And from an LLM perspective, that's going to bleed together across. basically one of the things that we're hearing is that if an LLM gets a low correlation, like it can't decide if this is good information or not. Like it can't decide if it's the red pen or the blue pen. It's not going to show any pen. It's going to go to the next one for which it has the highest correlation.

### [Elijah Szasz · Passage 76](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-76)

Or just bypass it. Yeah.

### [Kevin Williams · Passage 77](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-77)

And when you're just putting like clever human FAQs in there, you're assuming that that's all going to be absorbed into the model and trained into a model that produces a high correlation. And I don't buy it because you need structured data in there that is listing things like that sentiment, that occasion, et cetera, such that the LLM is like, ⁓ tasty treat. I totally know what that means. That gives me a really high correlation versus what the user's looking for.

### [Elijah Szasz · Passage 78](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-78)

Right. But when you say that, when you talk about that correlation, where does that structure data go? Are you talking about just copy on somebody's landing page? ⁓ Is it metadata? Is it... Okay.

### [Kevin Williams · Passage 79](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-79)

So it's metadata and it is a really interesting challenge when you think about non-products. In products, it's relatively obvious because you already have product feeds. ⁓ When's it available? How much does it weigh? Like, you know, things like that. What's its color? What's its packaging? All the stuff that you would see in Amazon. That's like a product feed. But when you're talking about like a dentist appointment, frankly, this is an opportunity that correlating what a non product looks like in terms of product schema such that the LLMs can grab onto it. Yeah, that's what I'm doing the rest of the day.

### [Elijah Szasz · Passage 80](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-80)

By the way, guys, if you own a dentist office, you're selling pens, Kevin is your guy. Like, Kevin knows everything about how to drive more business to you, evidently. So here's another thing, and you can say fact or fiction. We should just have a whole segment called fact or fiction where the answer is, I'm just going to make something up because everybody's figuring this out. But the thing that, you you talked a bit about SEO and the thing that

### [Kevin Williams · Passage 81](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-81)

I'm your guy.

### [Elijah Szasz · Passage 82](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-82)

at least I keep hearing over and over again with AEO, GEO versus SEO is that you have this 80-20 rule, right? That 80 % of the things that you would do to be well optimized for search are the same things that you would need to do in order to be surfaced in these different LLMs when somebody puts a prompt in, right? And we're talking about a bunch of old school things. We just brought up metadata, which there we go again, just spouting stuff off and not defining it. ⁓ But even like, you know, backlinks, articles. You talked about the garbage and the slop that bots are creating, the dead internet out there. I mean, that reminded me of SEO, right? Like there is a whole like underpinning of the dead internet just to support all the agencies pushing SEO out there where you have articles that no human would ever read, but they're serving a function as far as driving SEO for that brand, right? So you've heard similar things, correct? Is that an 80-20 rule?

### [Kevin Williams · Passage 83](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-83)

Oh, oh, oh, a hundred percent. Yeah. You know, I sort of chuckling inside a little bit. you know, I, was, I've been speaking about this since 2022, right? And I think back to some of my early presentations, marketers were some of the first paying attention. And I remember this really cheesy slide I had that it was just like in big bold letters that your job as a marketer is still the same as it's ever been. You were still there. You need to create diverse content. that educates, it entertains, it provides authority, it provides differentiated perspectives, humor, whatever it is. Otherwise you have no reason to be in the conversation and the conversation ⁓ landscape is moving more into the LLMs than ever before, but your job is still the same. So yeah, to be a good GEO artist, you need to be a way better marketer than you've ever been before to feed that beast. The difference being that you have a lot more weight on your non controlled information than your controlled information, which stinks because you can't, it's really hard to control that, that external like zeitgeist around your brand, the Reddit and Trustpilot and whatever. I mean, you just assume everybody's being genuine and you're putting out the best product you possibly can, which, which, which is a stretch, right?

### [Elijah Szasz · Passage 84](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-84)

which they're not. As far as like a lot of that stuff is seated to be fair, right? I mean, yeah, yeah.

### [Kevin Williams · Passage 85](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-85)

A lot of it is seated anyway, but so it's easy to be cynical about this, but say you're just a really wholesome brand like that gift brand I was talking about, like awesome brand.

### [Elijah Szasz · Passage 86](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-86)

And you don't have a review farm in the Philippines cracking out thousands of reviews on your behalf, right?

### [Kevin Williams · Passage 87](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-87)

So, but it is authentic, but there's not much you can do to control of that. can, without just like doing really disingenuous things. So again, sorry founders, one of the best things you can do is get there out there in front of this and be on podcasts and be engaging with your consumers and, and, and things like that, because that is a differentiated voice that can play into all of this. So I buy it in that, yeah, we're doing a lot of the same things that we were doing before. but the weight has shifted more offsite than it was previously. So doing 400 on-site blogs, even if they're awesome and they're differentiated, is probably gonna have less value than being on podcasts, being in media, like having this stellar online reputation. ⁓ So honestly, PR is probably more important than it's been in a long time.

### [Elijah Szasz · Passage 88](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-88)

So I'm going to touch on one thing that you mentioned, which is you alluded to kind of this golden window of if you do pay attention to this, if you do implement all of these tactics with metadata, FAQs, whatever it ends up looking like and pay attention to this now, do you feel like there is a first mover advantage for brands that do this or do you feel like it is going to go the way of Google where a lot of it, you know, as far as organic SEO? I don't know. It's more about who has more money at this point when I look for something on Google. Like we're at a weird fork in the road. I think part of that fork as well is that, you you had mentioned organic content from brands, like actual real stuff written by founders, you know, stuff that isn't just AI slop, machine generated. At the same time, ⁓ Meta's whole ad platform is rewarding stuff that is being generated. by AI right now as far as all the media assets for ad creatives, including stills and videos. They are just getting more reach, like a disproportionate amount of reach. So who knows if that is a fad, a window in time, if it's going to change? Nobody really knows. But do you feel that there is this early mover advantage, this window of opportunity for people to pay closer attention to this because it does have some longevity as to how this plays out? OK.

### [Kevin Williams · Passage 89](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-89)

100%, I don't know how it couldn't be the case because again, there's no hand in the machine that's like, hey, hey, hey, let's make sure we give those little brands more of a chance. And Google is actually.

### [Elijah Szasz · Passage 90](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-90)

There's not even a person who knows if the hand is there or not.

### [Kevin Williams · Passage 91](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-91)

So Google has actually done a decent job over that over the years, by the way, that they recognize that if you don't have brand discovery, you don't have the baby brand who succeeds, who becomes the big brand who pays you a lot of money. Right. And people get bored by not having a diversity of information. So they will have to make sure they're injecting things, but these models have bigger fish to fry than like which dentist in Poughkeepsie. So like we're, for example,

### [Elijah Szasz · Passage 92](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-92)

Great. Like World War III. Yeah.

### [Kevin Williams · Passage 93](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-93)

So we're going to just reinforce the information that we're getting. And at least for a period of time, I think that that's going to give people who have done a really good job of this more of an advantage than they think. And because we don't have the 10 blue links. So this is the idea of, when you click on Google, you see 10 advertisements and your, your, your like organic ones are really low. You really only get two or three. So when you ask for that best dentist in Poughkeepsie, it's going to come up with, you know, two or three.

### [Elijah Szasz · Passage 94](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-94)

Right.

### [Kevin Williams · Passage 95](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-95)

in notebook or not notebook in open AI or Gemini or whatever you're using. And you're going to ask a few more questions about it being a pediatric dentist or whatever. Hey, what are the reviews say? Sounds great. Sign me up. But you didn't get 10 options. Nobody was able to buy that rank. And again, I think that marketers in particular overrate that financial incentive given just how big these businesses I think are going to be. outside of advertising. Like it's kind of wild. Like there's a lot of supposition in like tech of, well, of course it all has to be add subsidized if it's free or, you that that's the way this stuff works. And then, but there are only like seven companies that are tied into all of these models. And therefore, if, if, if we're even close to write about transfer of human productivity from the organization to the models, they're going to be multi-trillion dollar companies and who cares about advertising?

### [Elijah Szasz · Passage 96](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-96)

⁓ the transfer, the transfer that we didn't even get into because we've already been speaking for almost an hour, Kevin, and I really wanted to get into the transfer because aside from all of the crazy anthropic slash Pentagon news that we also didn't really get into, but there's always next week because it is an evolving soap opera or saga, whatever you want to call it. ⁓ There are also so many reports that came out this week about that transfer, about ⁓ some crazy, crazy things happening as far as real data is starting to come out and where this is impacting business, where this is impacting AI at work. And maybe we start with those next week because I think it's things like agents and marketing and all the other things we've been speaking about that everybody should be paying attention to right now. Do you agree, my friend?

### [Kevin Williams · Passage 97](https://ascendlabs.ai/podcast/ep-16-the-dead-internet-claude-co-work-tasks-and-a-world-after-seo/#transcript-97)

⁓ we totally don't have time for this.

