Loading summary
Transcript44 lines
- [00:00]
Podcast Announcer
The Voices of Search Podcast is a proud member of the I Hear Everything Podcast Network. Looking to launch or scale your podcast, I Hear Everything delivers podcast production, growth and monetization solutions that transform your words into profit. Ready to give your brand a voice? Then visit iheareverything.com welcome to the Voices of Search Podcast. A member of the I Hear Everything Podcast Network, ready to expedite your company's organic growth efforts. Sit back, relax and get ready for your daily dose of search engine optimization wisdom. Here's today's host of the Voices of Search Podcast, Tyson Stockton.
- [00:43]
Tyson Stockton
Hey, what's going on? My name is Tyson from Pre Visible, and joining me today is Marie Haynes, who is the owner and SEO expert at Marie Haynes Consulting, which offers expert advice on Google algorithm updates and penalty recovery, helping businesses improve their online visibility and organic search performance.
- [01:02]
Podcast Announcer
This podcast is also sponsored by Ahrefs. What if I told you that you could monitor your website's SEO health backlinks and organic rankings at no costs? Sounds too good to be true. Well, it's not, because my friends at Ahrefs just launched Ahrefs Webmaster Tools. Ahrefs new Webmaster Tools product quickly helps you improve your site's visibility by pointing solutions to over a hundred technical issues that might be holding your search performance down. Plus, AWT monitors for backlinks so you'll know the most linked to pages and how those links are affecting your rankings. And AWT shares what keywords your website ranks for and compares how you stack up against competitors for key metrics like search volume, keyword difficulty, and traffic value. Look, monitoring your website used to require multiple expensive tools and now, thanks to Ahrefs, that's not the case anymore because AWT will help you monitor your SEO health, backlinks and keywords for free. And no, it's not one of those 14 day free trial offers. It's a powerful site audit tool that will keep working for you for free. So check out Ahrefs webmaster tools@ahrefs.com AWT that's a H R E F S.com.
- [02:19]
Tyson Stockton
AWT today, Marie and I are going to be talking about why LLMs are imperative to SEO. So with that, here's my conversation with Marie Haynes, Owner and SEO Expert at Marie Haynes Consulting, Inc. Marie, welcome to the podcast.
- [02:35]
Marie Haynes
Hey, thanks for having me.
- [02:36]
Tyson Stockton
It's. I feel like we're going back and forth a bit on getting you on, but I was excited to finally get a chance to sit down and talk to you. Been following your work for quite some time the podcast that you've been doing, also your newsletter. So really appreciate the contributions that you've been making to kind of the community over the years and maybe to kind of, before we get into the topic, news on your side, like, I know podcasts, there'd been a little bit of a break on it. What's kind of going on in your world?
- [03:06]
Marie Haynes
There's so much going on, I think I feel like the world is changing so rapidly with what all the things that AI is doing. I had a team for several years, a team of 12 that we did site audits and reviews. And in 2022, a lot of things happened that just caused me to retire my, my, my business and return to solo consulting. And at the same time, ChatGPT came out. It's hard to believe it's been two years now. Also, ChatGPT came out and Google came out with the helpful content update, which they told us was a machine learning update, a machine learning system. And I thought, well, I don't really know a lot about machine learning. So for the last two years, that's kind of been. My focus is understanding language models, understanding how AI is impacting our lives. And so I'm figuring out a bunch of things now. Returning to podcast, putting a lot more AI stuff in newsletter as well as the SEO stuff, and just talking about how the world is changing.
- [04:10]
Tyson Stockton
Excellent. And I mean, with kind of like that transition to how, like, has that been kind of a new fun challenge, like, going from having like, the group to being, you know, more of on the solo consulting side, like, how's that transition been?
- [04:26]
Marie Haynes
For sure. So here in Canada, when the pandemic hit, they forced us to go remote to work remote. And our office was this thriving office where there'd be a Google update, we'd get a case of beer and a pizza, and we'd all figure it out. And so going remote was really, really difficult. It kind of sucked some of the joy out of running a business for me. But then I think everything happens for a reason, because another thing was happening at the same time is that we're going to talk today about AI and about language models and maybe AI overviews, those obvious things that are happening in search. But what we didn't realize is that AI systems have been underpinning how Google ranks results for many years now. And a lot of the things that we were recommending as traditional SEO improvements were having fewer and fewer returns. So our advice, we could see that the websites that came to us, if you've been impacted by a Core update recently or in the last few years, really, it's because the AI systems are really figuring out what people want, what brands people want, what content actually is helpful, what content other people have found helpful. And so our advice was things that you can't fix with SEO was, you know, improve your reputation. I mean, you can do to some degree SEO can help that, but become known as the go to brand in your industry. Or instead of outsourcing and paying for low, low cost content, write, have experts write it, which is something that, you know, costs a lot more money. So, so it was a big transition for me. I found myself as I was doing consulting with, with sites that were struggling, saying it was basically grief counseling sessions because a lot of websites that thrived for years based on good SEO can't function in the way that Google has changed their search algorithms with AI. So I went through some difficult times. I did a lot of consultation for free with sites impacted by the helpful content system. And then it turns out that again, those are things that you can't fix with SEO. AI is helping Google do a number of things that as SEOs, we're used to optimizing for handwritten algorithms that can be reverse engineered. And now there still are things you can do. Like you can really understand your user's intent and be the best answer for them. But it's a lot, lot harder to rank now than. Do you find that. Do. Do you think that it's harder to rank now than it was years ago?
- [07:03]
Tyson Stockton
Yeah, I feel like you oftentimes get that question, like, is SEO dying? Like, where's kind of like bingoing. And I think it's becoming more complex. So it's like, it is harder in the sense that there's like the complexity. It's like you can't have kind of like the universal prescriptions as far as, like, well, I saw this over here on this website. So therefore I can apply the exact same strategy over here. And so I think like that layer of complexity has been increasing, but it's like, it still feels like a little bit like the game's the same. It's just our board and the pieces and things are changing, but we're still like, I don't know, like the underlining approach feels consistent to me.
- [07:44]
Marie Haynes
Yeah, I think a lot of, I mean, I think if we've been doing SEO right for, and you have a brand that people recognize and people you have, Google says in their helpful content documentation the importance of having an audience first. And that really spoke to me because the Sites that were coming to us with not doing well in Google's algorithms were mostly sites that only existed for SEO purposes. So if you're helping sites that legitimately have a brand and an audience, then not a whole lot has changed. You know, you should be creating great content and getting attention on that content. It's interesting to note that a number of sites that were impacted by the November core update, when we look at them, they have the type of link that used to work really well years ago. And a lot of these sites had really SEO built links. Like, there's no reason for that link to exist other than to try to manipulate rankings. And I've been saying this for years that with I think every Google update, the more that they use AI to understand what people like, the less that the traditional algorithms work. So I thought that was interesting. I think I always, when I do these episodes or podcasts, I like to talk about Nav Boost and the click signals that Google uses. So I don't know, can we go there? That's. I think that'd be absolutely. Okay.
- [09:11]
Tyson Stockton
Let's absolutely go there.
- [09:12]
Marie Haynes
So it's hard to grasp how Google could understand the helpfulness of content. There's a slide presentation from 2016 from an internal Google meeting that has the slide slide that says that we don't understand content. They just look at how people react to content and then use that. And so I think that's changed. I think that with the new versions of Gemini that underpin the machine learning that is used in search, I think they do understand a bit more about content today. But for many years, what we Learned in the DOJ vs Google trial, there's a Pandu Nayak testimony, which if you have not read that, I would really encourage you to do so. I've probably put over a hundred hours into reading this testimony, and I go back to it almost every day. They talk about Nav Boost and then they talk about three deep learning systems that are used in search. So one is RankBrain. We've heard of RankBrain. It's been around since 2015. The other is RankInBed Bert, which is. So you've probably heard of Bert. I didn't know when Bert came out that it's a language model, just like GPT is a language model, but it's a rank embed. BERT is a language model that is specifically meant to help Google understand the context of a query and then embed it into a vector space. So historically, I believe that what was in that vector space for websites was the Title of the page, maybe the headings on the page, and then information from the links pointing to the page, the anchor text tells Google a lot about what those pages are about. And then the other deep learning system is called DeepRank, and we don't know a whole lot about deep rank. But all three of these systems, it says in the testimony, are trained in part on qlik data. Now Qlik data, then they go on to explain that is more than just what people click on. And then we saw when the API files, the big leak came out, which those files have been on GitHub apparently for two years. But it was great that we, that somebody discovered them. You can see that there are attributes that show Google can look at. Now, we don't know whether these attributes are being used in search, but it would make sense whether you're the first click. So are you. If you do a search and somebody tends to click on your site first, that's a sign that people found it worthy of clicking on whether you're the longest click. And one of the things that I found correlated with pages that did well after core updates is average session duration. So having somebody having them stay on your site is usually a good thing. And whether you're the last longest click, which is interesting, that means you probably satisfied that search. And then the trial talks about other things, like in the Google results, what do people hover over? What do they like? Not even necessarily click on, but if you were in a carousel, maybe in an AI overview, and you were in the link carousel and people hovered over your result, it kind of indicates that they were reading it and that's some indication of potential helpfulness. So today all of those systems, they're trained on what people find relevant and helpful. And so we actually train the system. So what that means is our goal as SEOs should be to be the result that people click on and to be the most satisfying result out of all of the ones that exist, which is a hard thing to do.
- [12:40]
Tyson Stockton
I was having the same kind of like train of thought. I mean it was less focused on the three different systems, but I was like more thinking about navbus and then how that, you know. And also it's like for me kind of has been a little bit of like the light bulb of like, ah, yes, this. Now these things are starting to connect and make a bit more sense. It felt like a little bit of the, the validation of the missing piece.
- [13:04]
Marie Haynes
Yes.
- [13:05]
Tyson Stockton
But like it feels like as a practice or as an industry, we're almost like drifting back towards More of like PR principles around like awareness and recognition. And I think like another piece too that I've been kind of. And it's a little bit of a hypothesis, but I would love to get kind of your take on it is with the same elements that you described. I feel like from that there is a natural human bias to draft towards something that's familiar. And if I think about like, okay, we've seen studies showing that kind of like the bigger brands are capturing more and more of like the open search and traffic that's kind of leaving. You had ran study in I believe beginning of this year that was, you know, hey, it's not that less Google sending out less traffic than before. It's just because the bigger brands are capturing more. There's, you know, the space is getting tighter for the smaller players. I feel like that. And then when you hear from Google, it's like, okay, hey, they, they recognize that that is actually a problem that they, you know, are not happy with that but they're, you know, working through it. And Danny Sullivan, like one of the Bay Area meetups said that to the group was just like, hey, like I'm not happy about this, but it's something we are working on.
- [14:26]
Marie Haynes
Yeah, yeah.
- [14:27]
Tyson Stockton
To me it feels like a lot of that could maybe come back to that bias in like, yeah, just human kind of psychology of having more likelihood to click on something that they know versus they don't know. Like, I don't know thoughts on that. Like maybe too many assumptions in that line of thought or kind of what's your take?
- [14:47]
Marie Haynes
No, no, I agree with you. I think you're right. People have a bias. If I'm looking for, I don't know, a review of skin care products or something and I see articles from Vogue, from Self, from, you know, I'm probably going to navigate towards those. If in search I see an article from best skincare for you.com, you know, like something that is, I don't recognize internally, I think I'm more likely to see the answer I want on the brands that I recognize. But the problem with a system that is so dependent on clicks is that there could be new amazing content. What if that article on that website that I didn't recognize actually was like dramatically better than what was on the big brands? This is why I think Google is making some changes. And who knows, Just as we started recording this, the December core update launched and Google said another system is being updated. In those API files that we talked about, there's a newer version so we were all talking about version four and then there was version five and then version six came out, which I haven't seen anybody talk about. The difference between version 6 and version 4 and 5 is this one section and it talks about chunks, it calls it chunks. There's chunks of text. I think it might be related to passage ranking, but I think this is my theory. I don't know for sure. My theory is that the changes that Google made with Gemini last year when Gemini, or actually February of this year when Gemini 1.5 came out, Google said for the language model that it gave it a large context window. So I don't know if you've played with the chatbot, but if you use it, especially in AI Studio, you can put in like a million tokens of text. So that large context window, it makes sense that that also helps Google store more information for search. So I think that Google can understand, you know, when you index a page, when you ask for a page to be indexed, I think that Google vectorizes the page or tokenizes it and determines which chunks of content might be useful to show people. And so where in the past Google could store the page title, the headings, the schema, the maybe the entities within, I think Google can see much more. Now if I can take a language model, I can take ChatGPT or Gemini, look at a piece of content and say, is this thorough and helpful? Does it contain information that people are not likely to find elsewhere? Like, it wouldn't take long for ChatGPT to give me an answer that says, wow, this has got new, original, unique information on it. So I believe that Google is moving not completely away from click signals. They still want to use user satisfaction. But I do think that we're going to see a shift where truly original and helpful content has a chance to rank against those big brands. The problem with this is that it's really hard to create something truly original and helpful. So if you look at Google's documentation on creating helpful content, it says the very first point is, is it original? Does it contain original research, insights, analysis? And sometimes what I'll see is, especially in travel, well, people will complain that like the TripAdvisor and the, you know, well known hotels are, are outranking my post when I have original photos in it. If you look at it from the perspective of a searcher, it's the whole package, right? Like if I want to know, oh, I'm going to this one area of Mexico or whatever, and I want to see what I'm going to See, it might be a bit helpful to see an original photo, but you need more than that. You know, you need to be the person who, like, if you're actually a travel blogger who lives in Mexico or has traveled to Mexico, and you can write a blog post about hidden gems that nobody that most tourists don't know about, that's original content. But what we see is like, people are just writing the same thing that everybody else writes. And historically, if we could SEO that content, we could get links to it. It would convince Google's algorithms that it was better than all the other stuff that was basically the same. So the real key is in creating something new for the Internet, which is very hard to do if you're writing on a topic that's well known already.
- [19:09]
Tyson Stockton
Well, I guess in that same line of thought, do you feel as these systems and we can get into Gemini 2.0, but as these systems advance more and more, are we going to continue to see this kind of like, maybe diminishing weight of some of the legacy SEO factors? Because the systems don't need to rely as much on some of those things that they used in the early days because they didn't have this level of kind of like understanding of context?
- [19:39]
Marie Haynes
Yeah, I think it's already happening. I think with each update we see. So in the DOJ vs Google trial, in the Pandu Nayak testimony, it talks about how Google uses their traditional search algorithms to narrow down. I mean, if I search for something, there's trillions of results that could meet my need, and they narrow that down to hundreds, and then they take the top 20 to 30 results and rank Brain gives them all a new score. It re ranks the top 20 to 30 results. So that means in RankBrain, when Google first came out with Rank Brain, they told there was an article in Search Engine Land that they said, if you want to learn more. Actually, Danny Sullivan wrote it before he was with Google. He said Google told him that if you want to learn more about how Rank Brain works, we should be reading the word 2vec patent or paper that talks about how words can be vectorized and put into a vector space. So vector search is something that we need to pay attention to. But the word two VECT paper was 2013. It was the same time Hummingbird came out. Google was reorganizing the search engine to make it an engine that could use an understanding of language. But that was 11 years ago, almost 12 years ago now. And they've come a long way. So let's talk about Gemini 2.0, which as we're recording this was just announced yesterday and it will change the world in dramatic ways. So I don't know. Did you see the video of the guy going around London with his phone and he's riding on a bike? I'll explain it because people probably haven't, not everybody's seen it. And so he has Project Astra, which I got to demo Project Astra at Google. I o. It is basically your phone camera, sees things in real time and then you talk to your AI about it. And so he was riding his bike through London and he's like, oh, there's a park over there, what is that? And Gemini said, well it's this park. And he's like, can I take my bike in there? And. And Gemini's like, no, no, bikes aren't allowed in there, but you could go over here. And there's a few. You wanted a grocery store, there's grocery stores down the, down the street. And so he was talking in real time with an AI assistant that was seeing the world in real time. So that changes search because the things that are valuable to people aren't just information like text anymore. Gemini will, with 2.0. So Project Astra, I'll be able to, if I'm trying to fix my lawnmower, you know, I hold my phone up and go around and say like, you know, what's, what's wrong here? It's making this noise. It's like it's, you know, this part looks like it's cracked or something like that. Right. And so what's going to be helpful to people in that situation is probably not a big wall of text on how to start your lawnmower. In that situation somebody's going to want a video. And so AI allows Google to get better at understanding what's going to actually help people. And now we're taking search from a text based thing to the entire world thing. That vision is going to change things that sounds will be searchable. That's kind of hard to grasp. And I think we'll get to a point where this idea of like typing something on a keyboard and seeing 10 blue links is as archaic as the yellow pages. You know, some people still use them, but AI will answer, you know, all the text based questions unless it's brand new information that only an expert has. And, and then we'll refer people to the type of content that truly is going to be helpful to them. It's a, it's a whole different experience. And I don't Know if SEO is the right word like we still as long as there are businesses that need to be found, then there's need for people to optimize and to to understand how what it takes to get found. But I think a lot of what we do in the name of SEO today is dying is not really helpful anymore.
- [23:42]
Podcast Announcer
Time for a one minute break to hear from our sponsor, Pre Visible. So you're looking for SEO help and you got a couple of options. You could start replying to spam from agencies that claim they can get you to rank number one on Google. You can pay an hourly rate for a consultant who will inevitably nickel and dime you with hourly charges. Or you could work with a cookie cutter agency to quickly launch a strategy less project with low success rate. None of those sound very good now do they? Well that's where Pre Visible's integrated consulting model comes in. Pre Visible draws From a collective 40 years of SEO and digital marketing experience to unlock your organic growth opportunities. They build custom solutions that combine strategy, technical expertise, content and reporting to effectively operationalize SEO for your business. Pre Physical's four stage approach ensures that your SEO programs thrive by starting off with a strategy first approach. Then they support you in your efforts to create quality content, help you identify technical issues, and most importantly, they'll work with your cross functional teams to integrate your SEO strategies to make sure that your SEO budget actually drives results, not just your agency's bottom line. So join brands like Yelp, eBay, Canva, Atlassian Square, all who rely on the SEO consultants at Pre Visible. For more information go to Pre Visible IO. That's Pre Visible P R E V I S I b l e.IO.
- [25:10]
Tyson Stockton
Well I totally agree with that. And it's like almost like the discipline or the practice needs to be expanded to be more of like information retrieval, data retrieval systems. Because it's like those fundamentals kind of still exist in this. As you were saying that though it was kind of for me over the last few years I feel like there had been a decent amount of discussion around kind of like voice search and there's going to be this rise of voice search and kind of through that I kept kind of having like just a feeling of like what about like image? And I would see like some of the early things that Google will be putting out of like you know, how to fix the bike derailleur and they take the picture of the part and then it tells them it's a bike derailleur and whatnot. But I think like as you were Describing that like video has to be, I guess like one, the application of that makes so much sense. But then how the information's returned is probably also going to be starting to be like verticalized.
- [26:08]
Marie Haynes
Yes.
- [26:08]
Tyson Stockton
Or there might be times where the topic warrants you looking in a more traditional fashion at a web page or kind of like exploring through that lens. Or there's other times like in your example that it's like maybe an audio cue feedback is the right medium in other cases, maybe there's different ways of serving video back in that sense too. But I feel like to me it feels like we're entering more and more into this like verticalization of search where there may not be the one place that you go. It's more of like, oh, I have a question about like recipes, I'm going here, I have a question of about like I'm in shopping, I'm going like this direction. But like, and that's I guess just kind of personal assumption there. Like how are you seeing that? Like how do you see this evolving? Do you think Google will be able to keep kind of that all in one practice or do you see it kind of fracturing out?
- [27:07]
Marie Haynes
Well, so Google's mission has always been to organize the world's information. It's funny they say universally accessible to all. I think that websites are just a step in the road for Google. I personally believe that they've had this idea of an AI assistant right from the beginning or very soon after the beginning of Google. And so here's an example. Just yesterday again Google announced this thing called deep research. I don't know if you had a chance to use it. It's within Gemini Advanced and what you do. So this morning I'm trying to figure out what to get my children for Christmas. And they're getting harder and harder to buy for they're getting older and like, they don't want toys, they want electronic stuff. And so I had a conversation with Gemini where I said, you know, help me find something for a 16 year old within this price range. And what it did was it went off and browsed websites. It browsed 37 different websites, including Amazon, including articles on the best toys, articles on what's trendy this year for, for that age group. And then it produced a whole document for me that said, you know, if you're looking for something trendy, there's this, here's the clothes that are trending for that age group. And I didn't go to any of those websites. So now if I own those websites, would I be upset probably not because those websites were actually like selling the product, although some of them had lists with affiliate links. I think I'd be upset there, perhaps. So Google's goal is to find us the answer that satisfies our search and not necessarily to send us to a website. Whenever I bring this up, the thing that comes back is, well, Google makes their money from advertising. So how are they going to succeed if it becomes not profitable to produce content and be able to run ads, be able to do affiliate links, things like that? Well, there was a really interesting thing. There was a Gemini hackathon a few months ago where Sergey Brin was present at this hackathon. And he said, I don't know his exact quote, but somebody asked him about advertising. He said, I of all people am not concerned about the business model. He said, if you have something good enough with enough value, people will pay for it. And he talked about how people pay for ChatGPT. And now they had just come out with Gemini Advanced, which I know very few people who pay for Gemini Advanced, but people will because it's the deep research that I just did was worth the $26 a month that, that I paid for it. And so I really think that Google's business model is going to change. I think that advertising will change that the idea of a display ad or, you know, a blue link like click on our website because we have your answer is gone. Instead, what if, if somebody was searching for, you know, my site's been impacted by a Google update, instead of me running an ad for my website, what if I had a chatbot that was trained on all of my knowledge and documents that I put into it. I think I could see a notebook LM version of a chatbot that people could click on. And I pay for those leads because they're highly qualified. So the business model for Google is going to change, which means that what's valuable, what information is valuable is going to change as well.
- [30:33]
Tyson Stockton
On. I want to just hit on. I guess to some, maybe it'll sound like a minor point, but the, in a broader sense, the application of LLMs into a personal database. What I did see, and I forget how long back it was at this point, but you calling out how Gemini is now integrated with Google Drive. And so for us as being a business that uses Google Workshops space, that really caught my attention. And so that's been something that I like as a business. I don't feel like we've fully embraced it yet. And I've played around with different ways and that. But I feel like there's a ton of potential behind that because, you know, like, as a business, okay, we're going to store all of our client deliverables in their respective folders within these shared drives. Within that then you're going to have templates, you're going to have like all these different things of these like common tasks that we might do as a business. Then the ability to be like, hey, give me a rough draft of this based on these examples. Like, to me I feel like those are things that are seem small but like could really open up the world from like a stance of efficiency.
- [31:43]
Marie Haynes
Yeah, I feel like I haven't. We haven't seen the full benefit of what Google's going to allow us to do with our documents in Google Drive. Actually, just this morning I went into a document and it gave me access to Google vids, which I only played with it for a few minutes. But it looks amazing that you can take your documents. So say you write a report for your client and then turn it into a video. With Google provides you. They use AI to make the B roll, to make the title slides and then you can do a voiceover and like you could, you can make a video out of things very, very quickly. You could easily do it with Google Slides. I'm sure that that connects with vids as well. And then, yeah, I think that the day will come. Where have you used notebooklm?
- [32:32]
Tyson Stockton
I have not. No.
- [32:34]
Marie Haynes
Okay, go play with it after we do this call. Because at first I didn't quite get it, but then what I do is I put in. So let's say I want to write something about Google's algorithms. I put in the NYAC testimony. It's a Google product, it's totally free to use. You don't need to be paying for advance to use it. You can put in up to 50 documents or videos or you can paste text into it or websites. And then when you ask a question, it's not using. It's not AI making up an answer. It's actually pulling it from your sources. So I can be like, where did they talk about Rankbrain? And then it shows me exactly where in the sources that happens. So imagine I have all I have a notebook LM right now you can't share it outside of your organization. I think you can share it with like if you have a personal account, not a workspace account, you can share, you might be able to share a link, but workspace you can't. But it will come where. So I Might have a notebook for all of my clients, documentation everything we've talked about, all of our meeting calls and then instead of digging through like where did we have that conversation where we made the action plan for this? We can just ask our notebook LM to do that so everything that we have will be just so easily accessible and then turned into, I could be like, you know, create a slideshow for, to demonstrate how rankings have changed over the last five years. So there's, there's a lot that I think will come out of that. And then NotebookLM also has this really cool thing, this is the thing that everybody knows it for is the podcasters that you can put a document or multiple documents in and then ask for a podcast. And it sounds like two humans talking. So my favorite thing to do with this, I'm gonna go a little bit off topic, but it's such a cool thing to do is if you're trying to figure something out, like you're trying to figure out your business direction, your life direction, write it in a Google Doc or ramble it to a voice recorder. I use the Pixel recorder on my phone. Then I take that conversation. So I'm just like, oh, I think I want to do this, but I don't know about that and put that into 01, which is the new reasoning model in ChatGPT. So you need a ChatGPT plus account to do that and then converse with O1 and it's going to reason through and brainstorm on like, well, here's the pitfalls, here's some other ideas. And then you take that conversation and copy and paste it into NotebookLM and ask the podcasters to talk to you about your brainstorming ideas and they'll create a podcast with your ideas. You can ask for new ideas for innovation, innovative things. Tell me where, where what's good in this idea and what's not so good. And it's such a great brainstorming tool to do that.
- [35:15]
Tyson Stockton
That's, that is super interesting. I mean, I guess some thoughts I could see it being very confused by, but especially the notebook. LLM is something that I will absolutely be playing around with because I think too in the, in like I spent a good amount of my time on kind of like the hiring, recruiting side and also kind of like training and development and especially because so like pretty much everyone in SEO, we fell into this from some path or another. And then the start of your career is you just trying to, you know, scavenge around for, you know, who's authority, like who Actually is, you know. Right. Or who's just kind of spouting out stuff, but having something that you could also be aggregating all of your information to, to then make someone be able to find it. Paint things quicker, essentially.
- [36:07]
Marie Haynes
Exactly. Yeah, yeah, yeah.
- [36:08]
Tyson Stockton
Like, I've been working with one client, which is one of the SEO tool providers, on doing trainings for, like, their service team. And I've been working on doing kind of like different training documents. But I'm thinking now I can take my same videos and training documents and things like that, have them in a central home. So then instead of the user having to go, step one, step two, step three, hey, my client is in this situation, and then it would be able to service back something, you know, hopefully usable from that.
- [36:38]
Marie Haynes
Yeah. And what you could do is have all of your training documents in one notebook and then give the chat information about your specific client and say, make a version that's tailored to them. Like, I'm starting to do that with my newsletter. So my Marie's Notes, which is the paid version of my newsletter, I'll put all of it in. And then I have one client that's an E commerce client. And so I'll say, make a customized version of newsletter just for this client. And it just pulls out the times when I've talked about E commerce or things that would be relevant to them. And so I think we've just scratched the surface on what we can do there. But I think for anybody who has vast amounts of knowledge and documents and data and video, that it's incredible for hiring. I hear a lot of people saying, like, we don't want AI to make hiring decisions because there can be bias. There can be, you know, for whatever reason, like, you'd hate to be declined a job because some AI system decided you weren't good enough, but you could use it to help you organize. So if you put, you know, a bunch of resumes or applications into NotebookLM, you could ask NotebookLM, show me which ones you know, have this type of. Of education or have experience in WordPress or, you know, things like that. And then you can use it as a tool to help you where you make the ultimate decision at the end.
- [38:02]
Tyson Stockton
Well, I mean, I could keep going down this road for quite some time, but it feels like as we're discussing this too, and I mean, with the recent news where Gemini 2.0 and you kind of mentioned the O1 as well, having a focus on reasoning, it feels like also we're approaching kind of this verticalization of LLM models of specializations within this. How do you see this kind of developing into kind of this next year with like these different variants of the LLMs coming out?
- [38:32]
Marie Haynes
So both Sundar Pichai, the CEO of Google, and Sam Altman, the CEO of OpenAI, said just recently in talks that 2025 is the year of the agent. And an agent is. It's a little bit hard to wrap our heads around because it almost seems not possible. But Sundar Pichai said we'll be surprised Even in early 2025 what we can, we can do with these agents. So an agent is an AI system that can see a lot of information and can take actions on your behalf. So again, announced with the Gemini 2.0 announcement was something called Project Mariner. And I only played with this for a little bit, but it is wild. So it works in your Chrome browser. And so right now you can use it. If you go to aistudio.google.com There's a version that you can, you can use it and you can screen share with Gemini. And what happens is, so I opened up my website, started scrolling through it and just talking, just like talking to my computer, going like, what do you think of the nav? Or like, you think there's too much text on this area? Should I break it up with an image? And then Gemini comes back and talks to me about it. And so eventually that's an early version of an agent that is taking an. I mean, I'm actually doing the scrolling, but what's coming next? And they demoed this at IO in May of 2024. They said, imagine, you know, one of the worst things about shopping online is if you have to return something. So pretty soon your Gemini agent will, you can say like, I bought these shoes, I want to return them. It'll find the email, it'll find the return form, it'll go to the website, fill out the return form for you, print out a shipping label for you, and does all that menial work that you don't want to do. So that's very similar to the deep research that I just told you about that it went and it browsed all of these websites for me. And there's all sorts of implications here. So Google made in their blog post about it, they have tons of stuff about how they've been red teaming, doing, like trying to figure out all the safety issues because essentially AI can just take over our computers and do a lot of work for us, which sounds terrifying, doesn't it? But the way that Google does it is the human always seems to be in control. So, you know, hopefully that's the way I. We can talk about all sorts of scary things. With AI, there's no end of stories about how things could go wrong. I thoroughly believe that there are far more people working on AI for good than for bad. And so, you know, I wouldn't. I don't want to let fear get in the way of me understanding these models because they're the way of the future. You know, right now, it just seems like a toy. And if you read the media, the media is so anti AI for good reason, because if I want to know what happened in the Leafs game last night or whenever it was like the last Leafs game, I can go read it on a website. Or what I like to do is go to Grok. Have you used Grok in Twitter's AI?
- [41:41]
Tyson Stockton
I have not.
- [41:42]
Marie Haynes
Okay, so pay attention to Grok. I think it's actually a bigger competitor to Google than ChatGPT. So if I went to Grok right now and I asked, tell me about the last Leafs game, it'll summarize all the news from websites. I asked Grok, what index are you using? Because we know ChatGPT uses Bing, and Google obviously has their own index for Gemini. Grok said that X has its own search index. Where did that come from? And so it would give me all websites I could go to, but also tweets of people talking about what happened. And so Grok gives me an answer based on what the world's been saying on X about this. And it knows who are the people who talk about the Leafs, who's well known, who do people seem to trust with this information? And sometimes it'll say things like, you know, most people. People are saying this, and people. And then there's a few people that have concerns about this and like, some fringe thing. And it, like, really presents it as like, okay, well, here's all the information, and then it's up to me to decide what to consume. So my point is that what we hear about AI is always going to skew negative because we always fear what we don't know. And also, a lot of traditional sources of information are going to become less and less necessary now, which is not, you know, it's when we. I really believe that what we're going through right now is. Some people call it the fourth Industrial Revolution, a technological revolution. When the Industrial Revolution first happened, before that 90% of the world was employed in farming. And then, you know, look at us now, the world, the revolution brought new jobs. But it was difficult, I'm sure, when they went through the changes. So I think we're going to go through some difficult times as people learn how to adapt to how the world's changing.
- [43:35]
Tyson Stockton
I couldn't agree more on that. And I feel like it's clearly going in that direction. There's no way I individually am going to, I don't know, change that or even probably have even a significant influence to that. So I might as well learn what I can about it. And the more I understand it, also, the less scary it's going to be. But I was in an interview actually yesterday with a content writer, and I asked the content writer. So. And I kind of got the feeling. I was like, this individual is a content purist, where it's like, through and through, just passionate about writing. And I was asking them, how do you feel about the rise of generative AI LLMs? How do you perceive this? Is this a threat to you? And his response was very similar to what you're saying. And he called out too. He's like, yeah, actually, when, like, the printing press came out for the first time, there were a bunch of people that were upset and terrified by it. And then we saw what happened to that. The same thing happened with the typewriter and so on and so on. So I thought that was like an interesting kind of, like, anecdote of, you know, even within a kind of a niche that also is, I think, has a lot of pressure from these things. There's still that awkward optimism of, like, yeah, there's still, like, a level of, you know, originality that we have, and the creativity factor is the piece.
- [44:58]
Marie Haynes
Yep.
- [44:58]
Tyson Stockton
That kind of is so difficult.
- [45:00]
Marie Haynes
So Jensen, Jensen Wang, the CEO of Nvidia that makes, like, most of the chips that are used for AI, he's the one who said, AI is not going to take your job. Someone who knows how to use AI will take your job. So if I was a content writer right now, now, like, I don't like giving something to AI and letting it write. I do think that Gemini and ChatGPT as well, is getting better at not sounding like AI, but it still doesn't sound like my voice as much as I try to train it to sound like me. But I can't imagine writing without the help of my AI assistants, because I'll be writing an article and then I'll put it into one of the tools and say, is this accurate? You know, can you find problems with it? Can you make it more concise? Because I'm Always too wordy. I really like using using AI to help understand what is the likely intent of somebody who's reading this article and how can I do better at meeting that intent. So for a content writer, a lot of content writers are going to lose their jobs. It's already starting to happen. I know SEOs that used to farm work out to content writers that now are like, you know what, I can get a good first draft from ChatGPT and then I modify it to make it, you know, our brand voice. And so content writers are already losing jobs, but there will always be need for writers. So content writers that can learn how to write better and more insightful stuff and get right to the point and improve their writing with AI will be very, very valuable.
- [46:30]
Tyson Stockton
100%. And it's like it's those net new contributions to the community knowledge base, whatever terminology you kind of want to use in this sense. But I think like that also kind of brings it back to some of like where Google's rewarding things and there, I think in nature there has to be that somehow kind of like reward to the uniqueness, originality, like something to diversify the results. But I feel like I could keep going on and on about this topic with you. So it's been an absolute pleasure. But with that, that's going to be wrapping up this episode of the Voice of Search podcast. Thank you again for joining us. I look forward to hearing and kind of seeing you bring back the podcast a bit. But I'd say too, if anyone would like to get in touch to you, we have a link to your LinkedIn profile in their show notes and be sure to check out her website@mariehaines.com okay.
- [47:27]
Podcast Announcer
Thanks to Tyson Stockton, our guest host. If you'd like to get in touch with Tyson, you could find a link to his LinkedIn profile in our show. Not you can contact him on Twitter where his handle is TysonStockton. Or if your team is interested in SEO consulting or organizational education, you can always head to their company's website, which is previsible IO that's P R E V I S I B L E I O. And a special thanks to Ahrefs for sponsoring this podcast. Monitoring your website used to require multiple expensive tools, but that's not the case anymore. Thanks to Ahrefs because they just launched their Ahrefs Webmaster Tools product which monitors your SEO health, helps you keep track of your backlinks, and gives you the insight into what keywords are performing for free. So check out Ahrefs webmaster tools@ahrefs.comAWT that's Ahrefs a h r e f s.comAWT just one more link in our show Notes I'd like to tell you about if you didn't have a chance to take notes while you were listening to this podcast, head over to voicesofsearch.com where we have summaries of all of our episodes and contact information for our guests. You can also subscribe to our weekly newsletter, and you can even send us your topic suggestions or your marketing questions, which we'll answer live on our show. Of course, you can always reach out on social media. Our handle is voicesofsearch on LinkedIn, Twitter, Instagram, Facebook, or you can contact me directly. My handle is benjschapp B E N J S H A P and if you haven't subscribed yet and you want a daily stream of SEO and content marketing insights in your podcast feed, we're going to publish an episode every day during the work week. So hit that subscribe button in your podcast app and we'll be back in your feed tomorrow morning. All right, that's it for today. But until next time, remember the answers are always in the data.