AI and tools like ChatGPT may change the industry as we know it.

Many in the marketing and SEO industry have been voicing concerns about AI and how it may lead to the demise of SEO and search marketing as we know it. We chat with SEO expert Chris Boggs to get his take on AI and its role in SEO.

Chris' website:

https://www.webtrafficadvisors.com/

[00:00:00] Roy: Hey everybody. Welcome to Real Marketing. Today we're going to be talking to Chris Boggs. He's the founder of Web Traffic Advisors. Chris has been involved in the digital marketing world since 2000, starting in the insurance space and then with increasingly growing leadership roles across various agencies and SEM firms.

[00:00:19] Roy: Chris founded Web Advisors in 2014 and he advises businesses and agency clients helping them to improve their SEO, paid search, and paid social media. Chris also loves creating and delivering customized team training and if you are ever attending shows, I'm sure in conferences, I'm sure you have run into Chris at some point because he is known for hitting the speaking circuit and being at the various conferences and stuff in the industry.

[00:00:46] Roy: And as a former head of SEMPO a number of years ago when that was still a thing. Chris is a well-known figure in the industry and today we're going to be picking Chris's brain about AI, Chat GPT, and other things that are taking place in the search arena and what we should be looking for and as always, Chris, it's a pleasure to talk to you. I'm glad you could join us today. 

[00:01:14] Chris: Thank you, Roy. Thanks for inviting me. It's great to see you again and I'm excited to talk about this stuff. It's fresh and new and it reminds me of when you know, some of the search tools first came out to replace, Overture, and stuff like that, right?

[00:01:33] Chris: Like where we used to be working in spreadsheets and suddenly we have these things that you push a button, right? So it's a great, it's a great day to be alive, as they say, right?

[00:01:44] Roy: You know what it does remind me of that kind of nascent period of where things just happened so fast. Especially with ChatGPT and some of the AI tools right now, I feel like just in the last couple months they've had a huge impact on the industry and people are talking about 'em and some people are afraid and other people are excited but yeah, it's one, I think it's going to be revolutionary and it'll be an interesting ride for sure.

[00:02:11] Chris: It's funny you say some people are afraid and others are excited because I was recently at a conference in Austin, a pretty well-known conference, and if it's all right for me to say the name, I'm sure you can edit it out as Pubcon.

[00:02:26] Chris: But the founder Brett Tabke really took a great liking to this subject and has for a few months, and he prepared a wonderful keynote, which I'll probably allude to a few times during this conversation. One of the first things that he introduced to us when people came in literally to the conference last week in Austin, there were people that were worried like, is AI going to replace me?

[00:02:53] Chris: And he said that AI's not going to replace you, but a person who's using AI will replace you. I thought it was a great kind of summary of this whole topic. And if you're not listening to these kind of conversations and testing it or whatever, depending on your role you may be replaced for not embracing it.

[00:03:14] Roy: Yeah. It's never really no different to some extent than the onset of Photoshop back in the early days of graphic design when a lot of designers were using pen and paper and doing markup and pasteup for publishing. And then you had the revolution of Photoshop and Quirk Express and the various different tools that allowed, graphic designers and layout to do things and a lot of designers felt like they were going to be pushed out cause now there's these digital tools and I think it's a similar concept. It's no different than, having tools to assist you more and work more efficiently in certain ways.

[00:03:55] Chris:  Yeah. If you're like a bricklayer or whatever, right?

[00:03:33] Chris: Like you can charge, 10 times as much because you could say you'll handle an average wreck, right? Or you can choose to use tools and do it faster and charge - there's still, I think, a means, right? For the people that don't use AI. I don't want to make that statement that I, quoted from the conference, right?

[00:04:17] Chris: My absolute, I think there's plenty of people that still love Excel and count me in that crew, right? But at the same time, it represents a great opportunity for people to I think spend time on doing stuff that they want to do, like strategy and ideation, and then once the ideas are formed, use the tools to grow those legs for you and then work with it. In summary and we're going to talk about some of that, I think.

[00:04:53] Roy:  Yeah, and I also think there's two ways to look at it too. There's the production of content and then there's also how search engines work. Is chat GPT or some iteration of it going to change how Google and Bing work fundamentally? Are people going to really be clicking through to websites or are they going to be getting results of their query? 

[00:05:18] Chris: Yeah and that depends on obviously the pace at which the technology gains the trust of the user. We can all go out right now and make the tool mess up. One of the first things that I did honestly on the tool was I asked it who owns Twitter, cause it was recently after Musk bought it, and it like very certain told me that it was the prior owner, and then I've seen, it get cajoled into like telling you that yes, you're right, your wife is right about three plus eight or whatever, right?

[00:05:55] Chris: I don't know if that one was real, but I've seen a lot of great examples where if you trust a tool to make major decisions for you, that's your fault. I think even with Google, if you're a sophisticated searcher, you're not going to just take the first result or the first five.

[00:06:13] Chris: You're going to take, if it's an important decision, you're going to probably want multiple feel-goods about it and that's how you should treat this. I think a good example, like, I was at first really excited about the possibility of this for like tourism entities or areas, right?

[00:06:34] Chris: Because you can say, Hey Chat GPT, make me a tour of the five best places to do across two days in such and such county. Maryland, right? Or whatever. 

[00:06:45] Chris: And cause it came out and it's about out these great things. But the problem is, right? Is that it that again the databases, right? Cause certain counties and certain areas have been descended upon by more marketers that have created top 10 things to do list or whatever. So there's going to be some list that's basically the opinion of one person. There's going to be I think in a scenario like that, there could be problems too, right?

[00:07:14] Chris: Like you could count on the tool to set up a great itinerary for you, but if you don't double check it, two of the places might already be shut down, etc., etc., right? And that's again, where I think the limitations exist right now. Especially with Chat GPT three. Obviously with Chad GPT four, they're claiming it's going to be that much better, and then, Bard, obviously Google's holding off on that. 

[00:07:39] Chris: Especially with things with fan answers to some of the questions it's being given. I think, it's again, you can't over expect out of the tool, what it can't deliver. You still have to, I think, use your human decision-making tree to really finalize your processes for important things. That said, if I wanted to write an article on something, if it's a medical topic, I'd want the doctor to read it over, right? But if it's a topic about French onion soup then it's different. Maybe, who knows, it can be really cool. To me, that's the opportunity is more on the creative side and helping to support that. 

[00:08:29] Roy: Yeah and I think that's a great point. Is that you, a knowledgeable person is not trusting everything they see in the Google search results today. Even the Google business listings are often wrong about hours and whether a business is still active or closed or even where it's located on the map So just user beware and consumer beware, I guess is the point going to be true going forward, no matter the tool. I think there's been some, initially when Chat GPT came out, a lot of people were saying, Hey, everybody's going to write their blogs using this. Like it's going to, you're going to be able to crank out all sorts of blogs and you give it some prompts and it'll produce content.

[00:09:11] Roy: There was an initial feeling that Google was going to squash that and that AI-developed content was going to be frowned upon by Google. They've walked that back a little bit. What's your feeling about what the ultimate area, and what the result of that will be? Is Google going to have an algorithm that identifies generated content and viewed as spam? Or are they going to, join the crowd. 

[00:09:42] Chris: Right now, and I'm looking, I put in like on Twitter or Facebook or something. There's a dozen good AI content detectors out there, right? So you can't expect that Google doesn't have, at least, that those type in play plus whatever else, right?

[00:10:05] Chris: I would suspect that it could become part of the algorithm, but not to get into a deep SEO discussion on content, but EGAT has to do about whether the content does have EAT. Now, the first ‘E’ think is hard or the new ‘E’ right, experience is hard to replicate with Chad GPT, and that's where the human layer I think is required.

[00:10:30] Chris: And if you could spin the content that's generated for you, you could get some good results. There were certainly some tricks that were mentioned at the conference around trying to basically launder your money with your content, right? But launder it through some tools that are able to rewrite content.

[00:10:50] Chris: And then do a couple things here and there to make it look like it couldn't have been written by a machine. cause it's not perfect, right? There's, Copy Leaks was there, they have a tool. There's a number 0GPT content at scale. Sapling, GLTR, Originality.AI is a premium tool.

[00:11:17] Chris: There's a number of like lawn tools out there already, right? So I think as an agency you want to CYA and basically establish a process where if you're going to use it as part of your content development process, which I don't think should be bad because it can as I mentioned get to scale as long as they're still a human, but it's still, you're going to have to have that conversation with your clients. I know there's a lot of different discussions going on right now about labeling content. I think there's no problem with saying but would you label your content and say part of this content was developed using Google to research the topics, right?

[00:11:59] Chris: It's almost like Google's pushing you into wanting to label it, but I personally think that's too far. Unless you plan to like, have an article written, publish it totally as is, and publish 10 per day, right? , Hopefully, if Google's got done anything right, the current panda would catch and throw it right out.

[00:12:21] Chris:  So to me in a way it could represent a challenge to Google, right? Can or does it better the content? I think it's pretty easy. I'm sure you've played with it, but as you start to ask it for topics it's definitely a very, as an educated human that's seen web copy and written copy for years, it's very easy to see that it’s a pamphlet, right?

[00:12:47] Chris: Somewhat sophomore or whatever you could call it, even maybe like reading a machine. And maybe that's good. I'll give you a great use case for it. And then I'm curious as to your thoughts for this. But I had a client who is passionate and knowledgeable about the industry, but not a great writer.

[00:13:10] Chris: And so the process of editing their content for optimization was much heavier because there was also an editorial layer required, right? So what we've basically agreed with the client to test is doubling the output of the content by letting us pick the topics and write the first part.

[00:13:35] Chris: Give it a human layer, give it to them, and then they'll write it and put it in their voice. But to me, that's an example where it's the same work as we do as an agency, right? And we put human hours into investigating. And so it might come across as being non-human or anti-human, but there's still a human that's gotta prop this, right?

[00:13:55] Chris: So that's the key and you have to do that overlay of the BS detector, right? Even these machines, you have to operate them, right? So I don't think that this is at much putting human writers at risk there. I think if writers embrace this, they can increase their output. But I would caution that, if you're used to charging a certain amount of hours per piece of content, it might, the expectation might change at the agency level or at a client level in terms of how long it should take to research content and to create a relatively good list of outliers. I might get daggered by writers about that comment, but that's fine, from my opinion. Yeah, writers needs to humanize it and I'll always stand by that. 

[00:14:44] Roy: And it really is, I see similarities to, cause we do the same thing. We are producing content and we help clients produce content about topics which we don't really know anything about. We're just cleaning up the copy, making it, wordsmithing it a little bit. Bringing in a little bit more professionalism to the copy because maybe the client's not a great writer or maybe we interviewed them and then we're taking a transcript or something like that.

[00:15:10] Roy: The one interesting thing that I've seen, I don't think right away, at least with what I've seen the output from Chat GPT that anybody's going to be using it effectively to write blogs. Sure. Are they going to be able to produce like a whole bunch of 'em and they're going to not look like great quality?

[00:15:27] Roy: So if you're looking for volume, you can do volume. But what I found interesting is Shopify just came out with an AI widget. For their product copy. If you're a Shopify retailer, you can put in some key terms, choose the tone of voice, whether it's professional or expert, or it gives you five or six different options.

[00:15:50] Roy: And then it'll generate a copy based off of the product category and the prompts that you gave up. It's decent product copy and individual product pages don't typically rank that well. If that can speed up the copywriting for some companies have products that are essentially the same product. It's just slightly different because it's a different color, so the copy's not going to be that different. If that can speed up the product copywriting and creation so you can at least have something that's unique, I can see that being effective 

[00:16:32] Chris: A hundred percent. To me, like that's a green area where it can make a difference even though there's hundreds, dozens sometimes if you're lucky, of competitors out there with the same product essentially. If you can differentiate that page and differentiate the product copy. It'll be interesting, right?

[00:17:00] Chris: Because to me, the test then is like, how good is the link strength to those pages, right? Because really many people make those pages rank without differentiating the copy, right? Through the way that they handle their internal link and external link strength. There's some testing you could do there, definitely.

[00:17:21] Chris: That would see if that gives it, then another plus right in the eyes of the way that it's noticed; if it's useful. I think we're really, even with these machines, we're getting more and more to the point where the machines can tell whether or not it's human right. I think that once you start to get a thousand different pages that are optimized for Nike shocks that are blue and yellow shoelaces, they're going to start to look the same.

[00:17:54] Chris: So again it's going to go back to who has the better internal links, right? And that'll be it I think in an SEO sense. But like when you're looking for that blue Nike shocks with the yellow shoe license, once you land on that page, if it's within 20 bucks of what you're going to pay, are you going to read the text?

[00:18:17] Chris: I think there's a usefulness that has to come in there and not every product has that. Nike shocks is fairly cut and dry. Maybe you want a pair of snowshoes and then you gotta consider weight and type of terrain and all that. Then maybe there would be more options.

[00:18:35] Chris: So I think deciding where to invest in that kind of stuff would be the answer there. You know, it's the French onion soup, are they human? Corey deserves recency sort of thing that they, we talked about 10 years ago or something where, you know, if it's french onion soup recipe, man, how different is it going to be?

[00:18:54] Chris: That's Google's predicament with the French onion soup recipe of right. So if you're a pair of Nike shocks out there, it's going to be different unless you really figure out a way, right? Then that's maybe being some sort of a shoe store that figures out, shoes by EU, which they've all done, right?

[00:119:16] Chris: Like running versus cross-train route, but then takes it to another layer, right? And to the northeast or so on and you leverage ChatGPT to do that or help with that. Kind of, I think, what led with is that there are some tools that scripts and so forth that can help you with the prompting, I think and that's really important, right?

[00:19:40] Chris: There was a great example he gave of a riddle where the system couldn't figure out it was a riddle and didn't have the answer until he said, Hey, this is a riddle, and then it immediately said, oh, here's the answer, right? So if you know where we're at with ChatGPT three, I think will be easier as per se for maybe the less sophisticated user with four, but it's still going to require a common, like what will become what, we're all used to searching Google now or Bing, or whatever you use, you get pretty used to your types of strings and so forth. I think that's again, where the value of a tool that Chad GT is that every, whether you're a writer or a strategist for a campaign or even like an executive that needs to get an understanding of something. Like I've told a couple of executives about this tool and they've looked at and oh my God, I've been looking for the way to better explain monetization of tourism this or that.

[00:20:44] Chris: It's amazing the kind of answers that the tool gives you. Even at the executive level, if you're looking for something to simplify your life and to better understand why this score needs server boosts or all kinds of little different things like that where you may not understand something. Or that is something kids these days are talking about, even that a lot of it is pretty accurate, you gotta be careful, post 21’ right now but it's learning. I haven't tried, to be honest, I haven't tried Bing. I've read a lot about, it's supposed frailties and obvious, some pretty funny stories.

[00:21:30] Chris: But the difference between that is that it's strong AI versus weak AI; I think is the terminology used within the industry where machine learning chat G B T type stuff. It doesn't still build its own mind from that, whereas Bing’s that stronger AI that's trying to be an actual thinking entity. I'm excited to learn about that too. I do feel that inherently it's creepier obviously the stronger the strong AI side has been seen but also I think it'll be more difficult to leverage effectively from marketing day to day stuff. The stuff about the weak AI is stuff is that it is it's dependent on what you put into it.

[00:22:23] Chris: So as you get better at that, I think that no matter again, what role you're in marketing or sales or just business management or whatever, it can be very useful even in office management, right? Or figuring out an Excel-like problem, all these different things. The tool has proven to me very useful.

[00:22:45] Chris: I paid the 20 bucks a month and I really think it's a big difference too. Which it does. It is phenomenal and I only expect it to get better. 

[00:22:57] Roy: And I've seen interesting iterations of the whole AI too, like I said in Shopify with the product descriptions.

[00:23:06] Roy: But I also have an email client tool for Gmail that now has a summary function and so you can actually click on the widget and it will give you, if it's a very long email, it'll summarize it for you. So from an executive perspective, if I don't feel like reading through a string of 30 emails and want to figure out what's going on, it'll do it for me. That is true. Stuff like that, obviously is an efficiency. 

[00:23:32] Chris: I replied to your email about this with Chat GPT response.

[00:23:38] Roy: Yeah, that was really good. So for people watching I had invited Chris to the show and he actually used Chat GPT to create a letter in response saying that he'd be happy to come and it was quite clever. That was pretty funny and it was a well-crafted letter. 

[00:23:54] Chris: Yeah. Again, and you add a little bit of your own to it or whenever it is, right? And an interesting part that came up in a different session was no one ever really updates their terms and conditions and so forth and those boring pages.

[00:24:10] Chris: But it's actually pretty interesting some of the experiments that people have done by, they can update some of that, and still you don't want that ranking or whatever, but you might as well make sure that stuff’s zipped up. And if it takes just the process of putting it through there and them saying here's recommended updates, then why not?

[00:24:30] Chris: Stuff like basic legal ease and stuff like that. You can even check in there. Again, don't take this to the bank but it's a useful tool. 

[00:24:41] Roy: I've been experimenting too because coming from academia and stuff initially, and I tend to end up being in a former English major. I have been scolded in the past for writing it too high of a reading level. So for certain copy you want to aim at that sixth grade, fifth grade reading level. Unfortunately, that's more of a statement I think about our educational system. For a lot of ways, for a lot of writing you want to aim at that reading level for some of the stuff that we do.

[00:25:17] Roy: And it's, stuff like that is a good metric of am I writing at the right level? Should I scale back on some of the large words that I'm using? And can it do that for me? Can it summarize it for me at a more attainable reading? So that's useful. We've done some experiments with stuff too, just as far as like copy for Facebook ads and headlines and you can turn out 10 headline variations, but telling at the tone of voice that you want that seems pretty darn effective.

[00:25:53] Roy: Like obviously just like with any of the other stuff, you want to touch it up. What have you been experimenting with? What have you seen? Where it could be useful for you. Obviously not writing necessarily blogs, but is there stuff that you're using it for that helps to speed up your day?

[00:26:10] Chris: The same stuff. You get a banner and you're going to run it in ads, right?

[00:26:20] Chris: Whether it's Meta or potentially Facebook, or you want to create another responsive unit in Google or Bing, right? And you want to just try to further give the system options and maintain control over letting the system give you more new ads. So, I think that this is a great opportunity to help do that and help keep the system happy in a way that you're not, especially Facebook, right? Because I think that Meta you can definitely, for imagery and stuff gets staler more quickly. From a headline perspective, a hundred percent right. But I think also on the image side and I know this is where it gets into, as you and I were discussing, prior I'm a little more quirky about as it were stealing hours from artists, right? I'd rather buy stock imagery or whatever or use even free stock photos based on and then occasionally buy up with, I use Unsplash, and I occasionally buy up and you provide, but honestly, like if I'm writing an article and I want to show a bull market or a real estate investment and I want to show a…I think that it's, using some of the tools out there becomes fair place.

[00:27:56] Chris: So to me, it can help also because it would scale and improve your marketing assets and it becomes a business decision then again. And that's where artists are expensive. And so if potentially, I think one opportunity for artists to work together as it were, versus railing against, which I totally understand, it would be to like start working into your contract like if I create a base set of art and you choose to replicate it through AI, then it's going to cost you this much each time. And then that gives, I think it gives legs to the person's art. Like the person who did my logo eight years ago hasn't gotten any work out of it cause I don't do a lot of marketing or any but theoretically, like if I wanted to spin it off in AI, then I would feel fine because I already had it in agreement with my artist. 

[00:28:58] Chris: So that to me would be one area to consider. That has to, I think, still shape itself out because we've seen like the power of the art side of this. And the video side, geez. You're talking 10, 15 grand to develop a decent video to use in a like a regional tourism thing. Whereas it's less considerably to use AI and that's the scary part for artists out there. I understand but this is the time where again, I feel that the human side of it and the ideation of the art, and then why not just whoever it is that created the NFTs was making a bunch of money signed. Fixtures of turtles or whatever. Why not just embrace it? And that's my opinion on that. What do you think?

[00:29:52] Roy:  Yeah and I just saw an article the other day or it was talking about a photographer who took a series of photos for a company of people at the company and took a whole number of pictures of those people used that to feed into an AI tool so that he could place those people into different settings so he could say, push this person, this female model I want her in a lab coat.

[00:30:22] Roy: And it would essentially render that photo for him and very well, but he didn't take the picture, but it was based on his.

[00:30:30] Chris: That he had paid for that.

[00:30:32] Roy: Right or that he had taken and that way he's I don't have to go back there and shoot a photo again for a while. I can just, for this client, I can actually use the photos that I took previously. 

[00:30:43] Chris: But the model didn't say that they were able to adjust the pictures or, 

[00:30:48] Roy: Right. That there would have to be some sort of, a sign-off on the models per point that they would be able to use their image for that. Yeah, it was people that were-

[00:30:58] Chris: Sorry to cut you off but is that basically what happened or they haven't had any whip back on that on it? 

[00:31:04] Roy: They were employees of the company. Yeah, it was employees of the company that he was producing these images for. But that way he could just put them in different scenarios and stuff like that.

[00:31:14] Roy: I want this person outside, or I want this person in a meeting, or working in a lab or something like that, and he didn't have to go back and take that picture because he had enough of a library of their image that he could use the AI to do that. Yeah. I don't know how.

[00:31:31] Chris: It'd be cool to develop like a background music because I know that a problem that I've seen people have, especially with YouTube, is that, they're doing some walk down the street video and somebody happened to walk by someone that's playing Madonna in the background or whatever. It aches, man video, right? So if you could create, leverage the tool to go into it and help you and I think that YouTube already does it itself, right?

[00:31:59] Chris: They like basically offer a solution but I think it'll be cool to, I think, leverage the tool in that way, potentially. To help cover your derrière from a copyright perspective. Ironically, creating music on it, say, oh no, hopefully not stealing. That would be funny if it would steal it.

[00:32:23] Roy: Or use something that you know, that a company is going to license it to you and they want to promote a certain song or something, and you could, you throw into the. . .

[00:32:33] Chris: Yeah. I'm the server and in that solution.

[0032:35] Roy:  Yeah, for sure. And I think where it gets really scary is with the deep fake kind of videos. I don't know if you've ever seen that not Keanu TikTok or whatever. It's really creepy how much that actually looks like Keanu and those type of deep fakes are scary, but incredibly neat at the same time.

[00:32:57] Roy: There's one now I forget the name of it, but it will allow you to essentially choose what star you want to be so you can take a video of yourself and you can choose the face and the voice of the well-known celebrity that you want to be, and actually will do a deep fake of you as if you were that person.

[00:33:17] Chris: There's some amazing ones, right? Like Spin Rewriter, there's rewriter tools and stuff like that. The content that I saw like Brett present at Pubcon was amazing, right? This deep kind of fake stuff. Synthesia and Runway are a couple tools that are out that. The Synthesia AI avatar is the example to go look up and check out, but it's amazing.

[00:33:58] Chris: She just basically interchanges, you can type in whatever you want her to say, or him. Can pick up from hundreds of different personas, right? To deliver the message. You type it in and then while she's talking or he's talking, you could switch it from English to Spanish, and the lip movement changes and everything and it's really creepy, but it was amazing.

[00:34:21] Chris: But then on top of that, he showed like this movie scene basically that had, been developed and his prediction is that like a couple of years from now it is going to be, within a couple of years, it's going to be the first release of all AI generated movie. Again, speaking of sort of high-value talent like for a studio or whatever.

[00:34:52] Chris: If they're going to see an AI movie that's going to be just as good at with a Transformer-type theme or whatever as Transformers and cost them a fraction of what having all those actors cost, I hate to say it, but that's an area where I think that AI really it goes against artistry right across the board, I think.

[00:35:14] Chris: Writing is the, I think probably the best where writers could still be allied with it, but this is an area where really artists including actors they need to see what's coming up here. And be prepared.

[00:35:35] Roy: Even just on a small level, we use a tool called Script for editing podcasts and videos and stuff, and it will learn your voice and if you make a mistake in your video, actually…

[00:35:51] Chris: Great. So if I screw anything up, you're going to fix it for me. 

[00:35:54] Roy: I could, yeah. So if I called you Randy, instead of Chris, I could literally go in and it's text, it'll transcribe your text and I could say, oh, that's where I called them, Randy, and I can say, fix this, put in Chris, and it will do it in my voice and you can't tell the difference. 

[00:36:10] Chris: That's awesome. 

[00:36:12] Roy: And you can do your whole sentences if you needed to. And it does sound a little mechanical like it doesn't have the emotional intonation and stuff, but some of them do. But this one in particular… 

[00:36:22] Chris: It sounds like they're getting better by the day. That kind of stuff works.

[00:36:24] Roy: Yeah but for a couple of words, like you literally can't tell that it wasn't me saying it. So if I misspoke and said, it being instead of Google, I could change that and you would never know. So it's a little creepy. Very handy cause I speak all the time. 

[00:36:44] Chris: Yeah, imagine Imagination Engine, I think is another tool that's out there to check out AI test Kitchen.

[00:36:53] Roy: Yeah. It's just really a burgeoning industry right now.

[00:36:58] Chris: Anthropic.com is a startup co-founded by X OpenAI employees. They've raised about 700 million in funding. They've developed an AI system similar to open AI strategy v D. that appears to improve upon the original and key ways. So I haven't checked that out yet. 

[00:37:15] Roy: I'll have to check that out too.

[00:37:17] Roy: Yeah. Just going back touch on it and just real briefly, we had talked about, how some people are saying like, the whole interface of search, the search activity could change and Bing’s trying to get first to market with their version of their Chat GPT integration and kind of an intelligent search.

[00:37:38] Roy: There was some spin initially that people are like, why would Microsoft want to do that? Because it would potentially take away revenue, right? If you're giving Chat GPT results instead of clickable ads and links to websites, could it actually take away a revenue source? And there was some theorizing that Microsoft doesn't care because they just want to hurt Google's search business; that they're willing, they're not making a lot of money on it anyway. So if they can get out ahead of it, undermine Google's search business and take something of a lead, then they can figure out how to monetize it. 

[00:38:20] Chris: With Google Ads. 

[00:38:23] Roy: Yeah exactly. Do you see that kind of rivalry or do you think that's what they're doing or they're just trying to get some publicity out of it?

[00:38:31] Chris: I think it's there's always going to be some element of that probably in there somewhere. I think that we've seen it over the years where Microsoft has tried to take on Google so many times since Google's dominance, right? And they failed miserably with a few iterations. I think Bing has done a lot better than anything else has in the past.

[00:38:57] Chris: Their ad systems' getting better, even though it's in the back of Google. They're continuing to fight an uphill battle against them. Most people I think, are going to find that if they're trying to find a solution, if they're not doing well, searching in Google or whatever, they might jump over to Bing and try that.

[00:39:20] Chris: They might try just to try, but if they're used to the way that they get results. If it's a zero-click search, it's a zero-click search, right? I think if Bing’s going after it and Google's it's a zero-click search, we're going to lose a few of those but Google's not capitalizing on zero-click searches either.

[00:39:41] Chris: They'd much rather search an ad that has an ad or five in it personally, I think, or a search control. So to me, it's also doing it tree fall in the forest to that effect? Like the types of searches where there's no ads. I think Google could care less if someone's trying to figure out what the square root of 17 is or whatever that the Bing AI will answer quickly. 

[00:40:09] Roy: Yeah. Yeah. No, I tend to agree with you and I think it's become so ingrained in our culture. It's a verb, right? To Google something. I think it would take more than just a intelligent AI search interface to upend that whole market.

[00:40:30] Chris: Unless you're willing to tell Bing, hey, pick me the blue widget I want to buy, right? You're going to use, you're going to go through and you're going to hit that transactional term by Blue Ridge eventually, and you're going to head the product you want. And I don't think that Bing can offer that within its AI.

[00:40:51] Chris: Until you've done a thousand searches with it, if you're locked in or whatever. Eventually, I'm sure it could learn what blue widget  you'd pick. But if you went and decided you wanted to buy something, run an experiment, right? Go find the one you want to buy, hey, and then let AI buy one for you at Bing and see which one you end up liking more and do it at least five times so you get a decent sample.

[00:41:20] Roy: Yeah. Yeah, for sure. So are you worried about where it's going or are you excited or which part of that crowd are you?

[00:41:29] Chris:  I'm excited. I think I alluded to human laziness earlier. Right now, as the father of teenagers I've seen the impact this has had on productivity for some students, not my children of course, but I've heard the stories and teachers when they caught on to it and stuff like that. And it's just, again, in a way like do we get cranky and old because they have access to the internet? And we had to use encyclopedias originally?

[00:42:11] Chris: Or do we embrace it and say, this is great. You have access to information quicker, so you should be able to make a better decision. And ultimately I think that's better, right? The more tools you give someone. So I'm not scared of it because I think that, as humans will continue to master the machine learning side of it.

[00:42:29] Chris: I think the AI side is just going to keep screwing up and asking people to dump their lover and stuff like that, or whatever, it's doing weirdly over at Bing. And humans will continue getting fed up with that and until they can figure out a way to remove that sort of flippancy or whatever that you find, and I think that's why Google, this is not, I'm not an expert in this, but I've heard that's one of the reasons that they're delaying into releasing Bard is because they're worried right about the attitude side of it.

[00:43:05] Chris: And I think that's, we're all humans and our kids are humans and they're eventually going to figure out something that works with this, or they're going to stick to the weak AI versus strong AI in terms of running their daily lives or helping with their daily lives and easing their daily lives.

[00:43:25] Roy: Yeah, for sure. I think that, there's ways to use it intelligently and, hopefully, we have that way. One of the good quotes that I saw was that, keep in mind that these tools to date at least our language processing tools, and they may come back with some really good language processing and that's a little creepy and seems like it's intelligent but it's not really an intelligent tool. 

[00:43:54] Roy: It's not an intellect that is processing like what should it be answering? It's just giving you, based on the data that it has, it's an algorithm at the end of the day that's saying, these are probably the best combination of words that I should put in here.

[00:44:11] Roy: It's not evaluating it for truth or whether it should say it or, that kind of stuff. It's a very good language processing tool. Very interesting result, but it's not an intellect. 

[00:44:25] Chris: Well summarized. I agree. And it's, that's the difference. 

[00:44:30] Roy: All right. Chris, I really appreciate you taking your time out of your day and spending some time with us. And we'll have to keep an eye on developments around this and maybe circle back and talk about this again in the near. 

[00:44:46] Chris: Yes, sir. AI will summon us to speak for now.

[00:44:52] Roy: We'll just have it, plan our topic of conversation, and do some deep fakes. And I'll just do your part, put a deep fake of your face on it, and then… We'll have to really do this. 

[00:45:03] Chris: That’s right. You have plenty of words. I'm sure I've said a lazy brown fox by now. 

[00:45:09] Roy: Exactly. All right, sir. Thank you. And for everybody watching the video and listening to the podcast, we're going to put links to Chris's information in the description and if you haven't heard him speak or been able to follow him if you are at a conference that you know he is going to be at, I encourage you to go to one of his presentations or moderations. He's always has some good insight to share. So thanks again, Chris, for joining us and I really appreciate it. 

[00:45:39] Chris: Thank you, Roy. It's been great seeing you and having a great conversation about this stuff is really exciting.

[00:45:46] Roy: All right, we'll talk to you soon.

[00:45:48] Chris: You too.