00:00:01.06 evergrowmarketing That's hilarious. Uh, I started bringing this microphone to like business meetings, just meetings in general. When I worked at my, uh, full-time job and everyone was like, you should do a podcast. I'm like, well, let me tell you what. 00:00:15.76 Cody So what we're doing now. 00:00:17.41 evergrowmarketing Yeah. So, um, I wonder if this, when this YouTube video comes out, if my video is going to be choppy and laggy like it is right now. I don't know. The. 00:00:31.04 evergrowmarketing Elgato, if you're listening to this, your product is terrible. Your Cam Link product. I hate it. 00:00:36.36 Cody This is you, you're the one who figures this stuff out. I don't even know. I've got like the most vanilla ah set up. It's really just like the you Google the thing and then people say, yeah, this is what you should do. I think that's all that I have right now. And it's not exciting or spicy or anything, but. 00:00:53.23 evergrowmarketing Yeah. So I said, I recently got a new Ram. This is the, not the cold opener we're going for, but, uh, I recently got a new Ram and, uh, so now it's like all four slots are filled up with like the same RGB stuff. And I'm like, Oh, I just think fun with it. So this time when the computer is cool, the Ram is like turquoise. And then when it's like within like at like 40 degrees Celsius, that's white. And then when it's like 60 degrees Celsius, it's, it turns red, but it's not like, uh, 00:01:21.64 evergrowmarketing It doesn't like just switch. There's like a gradient that goes. So right now it's kind of like a, like a darker pink. And so I wonder if my PC is just getting hot and meets the cool down or something. 00:01:36.01 evergrowmarketing So. 00:01:36.11 Cody That's a question I had. Okay. Real cold open. John, are you techy? Do you care about Ram size and, and your whole desk set up? 00:01:46.15 Jon _Originality_ai_ So I'd say I'm very practical with it. like if so So I don't like geeking out on that stuff, but I do you want to have stuff that lets me be as productive as possible. So I'd say maybe a bit of a nerd on monitors. 00:01:58.85 Cody No. Okay. 00:02:01.18 evergrowmarketing I have what do you have? 00:02:01.27 Cody Monitors. 00:02:02.12 evergrowmarketing What do you have for a monitor right now? 00:02:04.00 Jon _Originality_ai_ So right now, my home setup is like my 15-inch Mac and then a 32-inch. curved monitor and then my setup at the office is two, two 32 inch curved. 00:02:15.12 evergrowmarketing Okay, that's That's what I had before I see with 30 intent 32 inch curve, but I had a 32 inch curved and then I had two 27 inch flat monitors. 00:02:29.42 evergrowmarketing So three. 00:02:29.47 Jon _Originality_ai_ Yeah. 00:02:31.04 evergrowmarketing And then I just switched to a 49 inch curved. 00:02:34.65 Jon _Originality_ai_ Okay, sweet. How do you like, how do you like solo single, single monitor? 00:02:36.22 evergrowmarketing So. 00:02:39.47 evergrowmarketing So it's not like, I mean, I like it better because it's cleaner and it's it's less cables. So I can have a big cable management freak, but. I can, it operates, you can set it to operate as two monitors. 00:02:51.18 evergrowmarketing So it splits down the middle and then I can like literally have like snap windows in the different sections. So I still have two. ah And I like it too, because when i when I do a lot of the podcast editing or video editing, I can take that window and I can stretch it across the entire 49 inches without any breaks or gaps. 00:03:03.27 Jon _Originality_ai_ Okay. Yep. 00:03:11.21 evergrowmarketing And then it makes it really easy to find different parts and drag things to certain areas. So that's great. Um, the complaints I have is it's too heavy. So like, if I want to like move it or like unbolted or whatever, it's it's very heavy. But, uh, I found that when I am forced to work on only one monitor, like on a laptop, like I'm on vacation or something, I actually get more focused. Cause I can't, I can't distract myself with whatever's on the second monitor. 00:03:44.29 Jon _Originality_ai_ Yeah, yeah, interesting. It's one of my like pet peeves with like an employee is like, when I see someone show up to um work with like, just their computer and like no external keyboard, no external monitor, it's like, you showing up, you showing up without all the right tools to like to to do the job at hand. 00:04:03.16 Cody Yeah, I get that. Yeah, it's a I've dealt with a lot of the the pain stuff that comes with it. 00:04:05.31 evergrowmarketing Yeah. 00:04:09.90 Cody I'm big on ergonomics and I constantly ask our team, like, are you good? You just making sure that there's nothing that you need because based on what I'm seeing here, this is going to hurt um a year from now. 00:04:21.25 Cody So. 00:04:21.64 evergrowmarketing Well, Cody's a different breed. Like everything hurts him. He's got glass bones and fragile skin. So like his, his, he's got like the left-handed track ball mouse and like, he's got the, he's using the split keyboard right now. 00:04:36.31 Cody Yep. Yeah, it's split. 00:04:37.06 evergrowmarketing Yeah. And so, uh, Cody's Cody has wrist issues. I just, yeah I've never had wrist issues. I don't know how, maybe he just write types more than me. 00:04:47.27 evergrowmarketing I don't know. He works more. 00:04:48.84 Jon _Originality_ai_ I have I have I've gone to like not not full split but they're like curved, curved keyboard, side, side mouse. 00:04:59.00 evergrowmarketing This is mine. and For my Game of Friends, I have a have a ducky. It's just a so ah flat keyboard that makes really nice clicking noises. 00:05:07.78 Jon _Originality_ai_ Nice. 00:05:08.19 Cody And say it gives you the sounds that you want to hear. Right. That's what there's a strange overlap between a very nomic crowd and then the people who just like the I never understood that where they're like, this has this. 00:05:12.14 evergrowmarketing Yes. 00:05:19.50 evergrowmarketing ah ASMR of, I think these are you think these are blue. 00:05:21.35 Cody Yeah, decibels and. 00:05:25.08 evergrowmarketing Uh, MX blues. So they make the, like the highest pitch clicking noise. 00:05:27.97 Cody Yeah. 00:05:30.88 Cody Yeah, I don't get it. 00:05:31.82 evergrowmarketing Uh, Oh wait. Okay. So original cold open. John, are you drinking anything? And if so, what are you drinking right now? 00:05:38.23 Jon _Originality_ai_ Uh, I'm not, i I did just finish a Diet Coke though. 00:05:42.65 evergrowmarketing um Okay. 00:05:42.67 Cody Ah. 00:05:43.23 evergrowmarketing Diet Coke slab. We just did a diet Coke, uh, tier list in our, our private chat at work. Uh, cause we had tried a Oreo Coke zero. 00:05:54.07 Jon _Originality_ai_ I'm interested. 00:05:54.50 Cody Oreo. 00:05:55.78 evergrowmarketing Yeah, it was Oreo flavored. 00:05:59.53 Jon _Originality_ai_ Yeah. 00:06:01.26 Cody It was exactly what you would think it would be. 00:06:03.54 evergrowmarketing Well, that's so hard because you're like, I don't know what I would think. How would you, how would you put the Oreo in the Coke? Like, and then, yeah, they said very, there's actually an Oreo aftertaste. 00:06:09.91 Cody But that's what it was, right? I mean, 00:06:13.99 evergrowmarketing It was weird. 00:06:15.34 Jon _Originality_ai_ Yeah. That one, I feel like, like chocolate milk or they're being like an Oreo milk. That, that would make a lot of sense. 00:06:20.48 evergrowmarketing Hmm. Yeah. It's protein shakes like that. 00:06:22.75 Jon _Originality_ai_ Um, is there? Yeah. 00:06:24.96 evergrowmarketing Oh yeah. Uh, I'm weird that my favorite protein shake is banana banana flavored protein. 00:06:26.01 Jon _Originality_ai_ makes 00:06:30.79 evergrowmarketing They taste like yellow Laffy Taffy's, which again is, I guess doesn't like help me in any case. Cause people, I guess hate the yellow ones, but okay. 00:06:40.01 Cody Are you are you a regular Diet Coke guy? Or is it just a coincidental day thing? 00:06:44.60 Jon _Originality_ai_ Yeah, I know it's a probably probably a habit I do too much. 00:06:44.65 Cody But 00:06:48.35 Jon _Originality_ai_ like and I'm not crazy. I don't know. One, two a day, pretty consistently. Wish that was zero, but I don't know. 00:06:57.83 evergrowmarketing Yeah. When I have them, that's usually what I do. 00:06:57.96 Jon _Originality_ai_ Works. 00:06:59.39 evergrowmarketing Like when I, if I have like a 12 pack, like that's what I'll do. 00:07:01.67 Jon _Originality_ai_ Yeah. 00:07:02.21 evergrowmarketing But if I don't have it, I don't go out of my way to get it. 00:07:05.25 Jon _Originality_ai_ Yeah. 00:07:05.39 evergrowmarketing So I might have, we might have Coke in the house for like a week and then they won't come back in the house for like two months. 00:07:11.86 Jon _Originality_ai_ Yeah, that's good. 00:07:13.99 evergrowmarketing So I'm drinking, um, my Huell daily greens, not sponsored, uh, just finished creatine and also my Menards cup. 00:07:27.61 evergrowmarketing coffee. 00:07:29.67 Jon _Originality_ai_ Sorry, full arsenal. 00:07:30.48 evergrowmarketing yeah I live in the entrepreneur bro life right now. 00:07:30.73 Cody Yeah. Yeah, I was going to say, Jake, he's really got the persona down with the I don't even know what's happening. He's drinking a green thing and it's good for his health or something. 00:07:40.92 evergrowmarketing I'm just just trying not to hate my life. 00:07:41.50 Cody And I'm over here. 00:07:47.52 evergrowmarketing What do you got Cody? 00:07:48.22 Cody Anyway, what? 00:07:49.82 evergrowmarketing What do you what are you drinking? 00:07:51.16 Cody Water. 00:07:52.52 evergrowmarketing Vanilla. 00:07:52.84 Cody Yeah, super, yeah super exciting over here, right? But. 00:07:57.08 evergrowmarketing Well, uh, okay, let's get into it. Where that's an F of the cold open. Uh, John, um, welcome to the agency growth podcast. It's good to have you. 00:08:07.98 Jon _Originality_ai_ Yeah, thanks for having me. 00:08:08.99 evergrowmarketing Uh, so just wanted to kind of give a brief introduction to our audience. Um, so we have John Gillam. Am I saying that right? 00:08:15.52 Jon _Originality_ai_ Yep. 00:08:16.83 evergrowmarketing Gillam. Okay, cool. Um, you are founder and CEO of originality AI. Um, found at originality dot.ai. So that's how you pronounce that brand name. 00:08:27.85 evergrowmarketing We had Tim Solo on from Ahrefs and nobody knows how to pronounce that company's name. It's always like Ahrefs or Ahrefs or off refs or whatever. 00:08:36.30 Jon _Originality_ai_ Yep. 00:08:37.24 evergrowmarketing So. 00:08:37.52 Cody um ah 00:08:38.20 evergrowmarketing ah yeah 00:08:39.60 Jon _Originality_ai_ Oh, ARFs. 00:08:41.86 evergrowmarketing So, um, I think, uh, these, these start out all the same when we have our guests, which is just kind of like, John, give us a full, give us your background, not just like originalities background, but let's, I want to know more about you and then why you're, I guess, kind of here today. 00:09:01.86 Jon _Originality_ai_ Yeah, it sounds good. Yeah, we, so my my background was started out, a mechanical engineer at school, worked at a hard tech startup, a small company, um this like 2006, seven timing. um And like, we need a website. Didn't know how to do that. Went on, oh, it was oDesk at the time to sort of like found this like global pool of talent. And I thought that was pretty cool um of building, building things. And then that sort of started the thread that I pulled on. 00:09:34.43 Jon _Originality_ai_ ah Since then, which was around building at the time, building content websites, ranking them on Google, um generating generating value, whether that was through an e-commerce site, whether that was i own my own content portfolio. Went off to work as a mechanical engineer, i always had the side hustle, always wanted to get back to my my hometown. My wife and I were from the same spot, wanted to sort of ski and bike with our with a family in our in our hometown. 00:10:02.13 Jon _Originality_ai_ um And sort of thought the online entrepreneurship path was the was the route to do that um pulled the pin on day job in like 2016 17 timing. One of the big drivers of that was we built up a content marketing agency that was where we sort of scaled the content writing service business where we were having working with um customers buying content from us, working with writers, and then sold that. 00:10:27.74 Jon _Originality_ai_ And then that's what sort of led to the two of these sort of concepts around originality, which was a complete QA QC tool for ah content aye teams. 00:10:38.56 Jon _Originality_ai_ The launched AI detection the weekend before chat GPT launched. And then it's been a bit of ride since then. 00:10:43.34 evergrowmarketing Ouch. 00:10:46.16 evergrowmarketing What, I want to hear more about your, yeah, I do too. 00:10:47.26 Cody I got questions. Oh man, that was fast. Um, I want to bring it back. You and build sites. 00:10:54.79 Jon _Originality_ai_ Yep. 00:10:57.16 Cody You're ranking sites. I assume you were monetizing this portfolio, this network of sites. 00:11:03.98 Jon _Originality_ai_ Yeah. Affiliate, digital ads, and some e-commerce. 00:11:08.28 Cody Okay. So you got them all covered. Great. I mean, that's, those are the big three. Um, was there one that you liked more than the others? 00:11:15.76 Jon _Originality_ai_ uh, affiliate, but always felt very transient and like we weren't adding enough value into the world with it. And so longer term e-commerce. 00:11:26.53 evergrowmarketing Just transient mean. 00:11:28.64 Jon _Originality_ai_ Like it just, maybe it's the right word, maybe it's the wrong word, but just like there for a little bit gone for, for a bit, like, you know, my, my best affiliate site in 2015 doesn't exist right now. 00:11:34.45 evergrowmarketing Okay. 00:11:41.43 evergrowmarketing Gotcha. 00:11:41.69 Cody Yeah. 00:11:42.16 evergrowmarketing Kind of like a, like, it just kind of exists to exist. It doesn't add like value and then yeah. 00:11:46.71 Jon _Originality_ai_ Yeah, I like certainly tried to to add value and like I could easily make an argument that it did, but you know in the end, I don't think so. And certainly in Google's eyes, it didn't. 00:11:56.06 Cody Well, probably especially with a portfolio of sites, right? I think the few times that I've seen it work really well for people where they're still with it. They're they're deep into something, right? 00:12:08.03 Jon _Originality_ai_ Yeah. 00:12:08.50 Cody I think the one that comes to mind is like the um what is sleep like the dead. That's the one about mattress and those sorts of reviews. Seems like they're deep into it. 00:12:20.11 Cody um But you got to really love something. 00:12:20.47 Jon _Originality_ai_ Yep. 00:12:23.62 Cody I don't know. 00:12:23.91 Jon _Originality_ai_ Yeah. 00:12:24.53 Cody I don't know if there's anything that I love um enough to to keep going down that route. But so you from that, you went on to the next evolution, I guess, which was content, yeah publishing, selling articles, what'd that look like? 00:12:41.32 Jon _Originality_ai_ Yeah, so it we sort of. Based on sort of the engineering background and then worked in project management, I always sort of thought that like my sort of unique angle on online entrepreneurship was going to be focused around so like building a team and building systems. And so we built up extra capacity within our content marketing sort of side of the business and then built it. So it started selling that and excess capacity as a service. And so people that wanted to come and have, hey, I want for a post on my local business. I want, I got five clients for my, for my agency that they, I want to do the page traffic, but I need somebody to plug in the, the content. And so we were, so they, a plug and play content marketing solution where we tried to go right from not just sort of like. 00:13:25.38 Jon _Originality_ai_ Content mail, here's the keyword, give me some some words, but do the full sort of like content strategy, content creation, optimization, QA, QC step, and publishing, and try and be that sort of full full solution. And and yeah, built built that up and then eventually sold it. 00:13:43.68 evergrowmarketing What was your acquisition strategy for those content creation clients? 00:13:48.48 Jon _Originality_ai_ um Organic traffic was was a significant part, so we're executing our own content strategy to to get them. 00:14:00.83 Jon _Originality_ai_ Specific networks, I think, had a small audience in the world of building building and ranking websites um that that helped get some some clients. And then affiliate, those would be probably the three, sort of some 00:14:17.26 Jon _Originality_ai_ High touch personal connections, uh, that were affiliates. Um, and then sort of some, some affiliates and then organic. 00:14:26.37 evergrowmarketing So this is like 2014-15ish. 00:14:29.78 Jon _Originality_ai_ 20, yeah, 20, 2016 to 2020. 00:14:34.83 evergrowmarketing and So Jasper was a thing. AI content was a thing then. Were you using that in your your production? 00:14:40.43 Jon _Originality_ai_ Yeah. We. We were one of the, one of the heaviest users of Jasper. um I think at one point they had a call with the CEO, Dave, and said that we were the, the he at the time, the heaviest users of the of their tool, ah which was in this design. 00:14:53.15 Cody It's gonna feel good. 00:14:54.30 Jon _Originality_ai_ Yeah. It was like, yeah, it's been a couple, a couple of tools that we were the sort of biggest abusers of some or heaviest users of some might say, but yeah, no, it's a, we, so we had a service and this is kind of what started us to see the need around originality. 00:15:10.53 Jon _Originality_ai_ But we had a service where we were selling transparent AI content. And so where it's like, hey, this AI content, your efficiency savings are getting passed to you, but we're gonna have a human a human in the loop to do fact-tracking and editing. And that was that that had a lot of a lot of traction very quickly with with that service. And then that was kind of what started us seeing, because questions came up around like, how do you know, how do you know your other writers aren't using AI? It's like, well, we have a policy, but with no controls in place, um you know, a lot of people are happy to pay $100, $1,000, whatever the rate might be for an article aren't super happy to find out it was copied and pasted out of chat GPT. 00:15:55.20 evergrowmarketing I hate that this is like the perfect transition and like the actual topic because I have different questions. So I want to put a pin on that and say, uh, and ask you yeah how many, how many content writing clients did you have and what size was the agency in terms of revenue but before you sold it? 00:16:02.20 Jon _Originality_ai_ Yeah. 00:16:14.91 Jon _Originality_ai_ Yeah. Um, two million a year revenue, uh, just, just under that. Um, and then number of clients, probably about 80, uh, clients could be, could be wrong, uh, on, on that. 00:16:29.15 evergrowmarketing Wow. 00:16:31.76 evergrowmarketing Okay. So yeah, it seems like you had, yeah, relatively similar prices to us. Then I assume these clients were like, you had like a reoccurring structure on producing content. 00:16:40.57 Jon _Originality_ai_ Yeah. 00:16:41.76 evergrowmarketing Okay. 00:16:42.06 Cody Well, but we're talking about content writing. So this is, I'm guessing this is pretty heavily profit, right? Are your expenses that high? 00:16:49.25 Jon _Originality_ai_ Yeah, so um i mean heart heart like we did a bunch of stuff to try and differentiate um in in terms of like added like value added. But in the end, hard the the majority of the value is a little word that are being produced. um Hard to sort of like put a markup on that. That is like 2x, 3x, 4x, what you're paying the writer. And so we were like a 30% net margin business. 00:17:17.68 Jon _Originality_ai_ to where where were goingnna aim to be off often fell a little bit under that, like kind of fell mostly around 20%, which was, I think for sort of i a business that has pressures like that, it was was okay. 00:17:32.12 evergrowmarketing Did you offshore a lot of your writers or did you have them locally? 00:17:36.58 Jon _Originality_ai_ All, with the exception of the AI business where that was a sort of like, we communicated that all native English. 00:17:46.45 evergrowmarketing Okay. Cool. 00:17:48.00 Cody Did, why did you sell? Why did you want to sell? 00:17:52.28 Jon _Originality_ai_ Yeah. um know So I had a great had a great team that was running it. 00:17:52.76 Cody You get tired. 00:17:57.91 Jon _Originality_ai_ um had Yeah, I had a great team that was running it. I had a like business manager that had that that was running it. 00:18:06.75 Jon _Originality_ai_ Right write price. I think i was I was overexposed on a few. like I had some other businesses. like We have a brokerage where we buy and sell websites in the marketplace that lets people list sites. 00:18:19.05 Jon _Originality_ai_ A lot of like my own portfolio that marketplace and this content agency all serve kind of the same audience around um people that were building sites, um small sites, small sort of entrepreneurs, not always small, there's some that were like million dollar sites, multimillion dollar sites, but very similar sort of like ranking Google, very similar business model around great content that ranks in Google and arbitrating it in some in some way. And so it felt overexposed on on that business model, despite having sort of a diversified portfolio of sites, marketplace, content agency. And that was and then the price was price was was right. 00:19:01.07 evergrowmarketing Do so. 00:19:01.14 Cody So partly you're just mitigating risk in your personal life. I mean, you just thought you were, you had too many eggs in one basket, sort of deal. 00:19:04.74 Jon _Originality_ai_ it 00:19:09.17 Jon _Originality_ai_ Yeah. Yeah. I think when i when I did sort of like a risk profile at at the sort of the the businesses, it was, it was clear that there was a few sort of events, Google updates, um, et cetera, that could, that could, despite them being what looked like different businesses, um, had some of the same, some same risk profiles. 00:19:27.65 Cody It's a very engineering perspective. I like it because you have, you have the other crowd or the other side who was saying, great. 00:19:30.32 Jon _Originality_ai_ Yeah. 00:19:35.30 Cody Well, if you're crushing it, then double down and shove all your money into it as, you know, as much as you can and wait for it to all blow up and have a glorious time when it does. Right. 00:19:45.80 Jon _Originality_ai_ Yeah, I mean, I think and it depends on the the side that you're coming at it from, from a business standpoint. Like I think a lot of, a lot of my focus around the business has been on other objectives in life around like being able to you know ski coach, sock coach, spend time with the family. 00:20:02.22 Jon _Originality_ai_ And so how do I sort of achieve, which isn't which isn't an insignificant amount, skiing's stupid expensive, but how do I sort of achieve what needs to be achieved with the lowest risk profile to to to me and the family? And so I think that risk has has always been ah risk mitigation has always been a sort of ah in the equation, not just the how do I get that number as big and fast and exciting as possible. 00:20:27.99 Cody That's awesome. I love that perspective. 00:20:30.85 evergrowmarketing I feel like skiing shouldn't be expensive. 00:20:35.95 Jon _Originality_ai_ it's 00:20:37.14 evergrowmarketing Once you have the stuff, i mean you just gotta find a hill, right? 00:20:39.24 Cody What's the expensive part? Yeah. Is it that you have to live next to the Hill or is the Hill itself the expensive? 00:20:48.05 Jon _Originality_ai_ So it depends. So racing changes the game because then the hill is approximately to the hill. And then the amount of capital investment that needs to go into the hills drives tickets prices way up. 00:21:02.75 Jon _Originality_ai_ And then the market has moved to trying to cater to the ultra wealthy as opposed to make it. So it's like fancy chairlifts are becoming a thing as opposed to just barely safe enough yeah 00:21:15.93 evergrowmarketing What's a fancy chairlift? 00:21:17.95 Jon _Originality_ai_ Heated seats. 00:21:19.04 evergrowmarketing well 00:21:19.17 Cody Why? 00:21:20.03 Jon _Originality_ai_ Yeah. 00:21:21.22 evergrowmarketing This is a world I've never been introduced to. 00:21:24.87 Jon _Originality_ai_ Yeah. And so it's this race, race to the top, like, like, I mean, I have a, I have a little bit unrelated, but I think this sort of like the, the barely good enough. So you can't buy a work truck anymore. Total, totally sort of off topic, but we have a bottom pool business ah locally. We need to buy a new truck and. 00:21:42.69 Jon _Originality_ai_ They're like sending like, Hey, just give approval for this purchase. And I'm like, why do we need like this size monitor inside of our work truck? And that's like, that's the base truck. 00:21:53.70 Jon _Originality_ai_ It's like every that sort of base expectations just continue to get raised. 00:21:55.74 evergrowmarketing Hmm. 00:21:57.71 Jon _Originality_ai_ Um, because there's just a lot of money to be made by, by, by doing so. 00:22:03.52 evergrowmarketing Yeah. Hmm. 00:22:05.84 Cody heated, heated chairlift thing, but my brain is still stuck on it. Like you, it's, it's snowing out there. 00:22:09.77 evergrowmarketing I I I was on like the yeah, I was on like the like the encapsulated lift, you know, like like ah we were just watching were rewatching the Girls Next Door series from the early 2000s with the with Hugh Hefner's girlfriends. 00:22:16.87 Jon _Originality_ai_ Yep. 00:22:25.31 evergrowmarketing And they were somewhere in Aspen that had that, they had like the gondola that, that took you to the top of the hill, but it was like a fully encased, like encapsulated ski lift. It is wild. 00:22:39.13 Cody Just shake your head. Like what a life. What a different life. 00:22:43.57 evergrowmarketing Yeah. So, uh, cool. Uh, sorry, that was a little bit of a detract from, uh, from your, I just, I always think it's neat when, um, People come on the podcast who own a a product or a ah software that agencies use and people would use, but don't realize that ah the the majority of the people that we and ah end up interviewing on here that come from some kind of like agency vendor perspective have previously owned agencies in the past. And so I think there's kind of like this veil with a lot of a lot of these companies, this but yours included that, 00:23:23.51 evergrowmarketing Some people might think just without even thinking that, you know, you might be out of touch because you are just a software company that is trying to sell things to agencies. But and on on the contrary, you know, you were one, a content one at that. And you developed this tool to essentially, you know, kind of seems like ah help your own case. But 00:23:48.33 Jon _Originality_ai_ Yeah. So we, we'd sold for sure. We'd sold the business, but, um, and then we, and we see it now with like AI detection. companies, they're they're all focused on very different things. but We're focused very much on the world of any company that has a copy editor. Tons of agencies have have copy editors. So we're not necessarily as focused on like, hey, we're selling to to agencies, although we have a ton of agencies. It's anybody that functions as a copy editor is a is it's kind of who we're who we're building for us. But yeah, I think there's no way we would have some of the insights that we have and some of the features that we've built that we've 00:24:25.46 Jon _Originality_ai_ made the right call on um without like we're the first ones to combine AI detection and plagiarism checking, which is just an obvious sort of combination now. and But yeah, I think it'd be really, really hard to not to to not have come from the world and then build a solution for that world that you come from. 00:24:49.24 evergrowmarketing I'm going to throw the i'm gonna throw there's the like the main thing out there because when we were approached with you haven coming on the podcast, I had heard of originality AI, but our agency doesn't use AI content, to my knowledge, asterisk. 00:25:06.51 Jon _Originality_ai_ Yep. 00:25:06.58 evergrowmarketing and ah When Chad GPT first came out and like and got big in like December of 2022, that was kind of like, whoa, this is crazy. 00:25:19.14 evergrowmarketing Uh, but we still hadn't really explored it and we still really haven't, but there are some Facebook groups, some large ones that were kind of diving into that. And I think early on. AI detection was pretty big. Um, and Google sent mixed signals about how they felt about AI generated content. Um, it's a little more clear on how they feel, which is that they don't really care as long as the content is helpful. 00:25:46.53 evergrowmarketing But whether people believe that or not is ah is a whole different story. um However, I remember originality AI coming up in these discussions in these groups ah quite a lot. And it was like like, as like the main AI detection tool that was being used. However, I hadn't heard a lot from it recently. And what I also knew about ah AI detection tools was that they they just seemed all over the place. Like they were detecting false positives, like human written content. They were saying it was AI, AI content they were saying was human. 00:26:20.87 evergrowmarketing ah And when I, I recently posted in one of these groups, has anybody heard about originality AI? And man, there wasn't like a single good comment. I wanted to be like, ah, this guy gets it. But like, everyone was like, Nope, doesn't do anything. Well, like it again, I don't know how they're using it, but I think it's important that I say that because I want. 00:26:46.17 evergrowmarketing it It feels like at a certain capacity, people aren't necessarily using AI detectors correctly, um, or responsibly, but I want to, I want you to kind of just go over, okay, what is originality AI? What's its purpose? And then how talk about the inception into where it's at now. 00:27:07.96 Jon _Originality_ai_ Yeah, yeah, so a lot, a lot to unpack there. um I'd say, i say i'm I'm, you know, there's, Google's got to say one thing. They don't necessarily always do what they say. um I think it's pretty clear that they are attacking AI spam. I would agree that not all AI is spam, but I think in 2020, 00:27:29.72 Jon _Originality_ai_ 24 all spam is AI. And I think it's a pretty quick mental sort of exercise to think if search results were filled with nothing but AI content, why would people go to the search results? And the sort of idea that you're adding information gain using an LLM because your prompt really good is a, is a leap. Um, that, that, you know, s it's an attractive leap because you're paying pennies when instead of hundreds of dollars for a piece of content, 00:27:59.87 Jon _Originality_ai_ And then you want to work because boy, I can just press a button and make a bunch of money. And that's an incentive. That's an attractive thing to want to be true. But I think what we have seen out of Google is that they have aggressively attacked anything that they sniff as AI spam. Their March 5th update, majority of sites that were manually de-indexed had the majority of their content was was AI generated. 00:28:28.89 Jon _Originality_ai_ um So I think that like that sort of the, yeah, so the the sort of last, it was a core update, but that involved a helpful content update um and the merch they did of manual action. 00:28:28.97 evergrowmarketing This is the helpful content update. 00:28:41.00 Jon _Originality_ai_ So it was, I call it like a, I think it was more of a PsyOps for them where, so, so, 00:28:44.84 evergrowmarketing Yeah, was it a manual action or was it just d like mass d-index? Because I think those are two different things, right? 00:28:53.53 Jon _Originality_ai_ They applied, it was a meat mass de-indexing through a menu by doing a manual action. 00:28:59.55 evergrowmarketing Okay. 00:29:00.54 Cody I like the psyops is a good. 00:29:00.90 Jon _Originality_ai_ And. 00:29:02.54 Cody That's a good combination of of words to ah describe what goes down when they they do these things. 00:29:05.44 Jon _Originality_ai_ Yeah. 00:29:12.09 Jon _Originality_ai_ Well, especially that one where they did it, they did, it was the biggest PR push where they're like, we're going to get rid of 40% of the spam on, on the web. And then they did this massive manual de-indexing on the evening to of the update where they talked about low quality content, spammy content, veiled AI spam content. And then they took manual action against. 00:29:36.46 Jon _Originality_ai_ tons of sites the majority of those sites that had the menial action had massively public, had a large rate of publishing of AI generated content. And the sort of PsyOps component was this like big PR push. 00:29:47.31 Jon _Originality_ai_ this manual action on the eve of the update, and then the update, and it sort of attempted, and I think successfully so, created this wave of like, oh crap, Google is slapping our w risk for publishing AI spam. So so that's the Google, kind of the Google piece to that statement of like, because otherwise, why would we care? Like just, why why would we care if someone was using AI content or not, if we weren't concerned that publishing AI content was riskier than publishing human content? 00:30:15.25 evergrowmarketing So I was actually, it was actually just on, uh, the SERPs up podcast with, um, Marty Oberstein over at Wix. Um, but yes, yeah, dude came over from SEMrush. But, uh, one of the questions he asked me was like, um, like what, like we're trying, we were talking about like moving away from like algorithmic checklists and building content and, and, and writing that for users versus algorithms. 00:30:41.53 Jon _Originality_ai_ Yep. 00:30:41.59 evergrowmarketing And one of the big arguments that I made was that. i don't I don't really like catering to algorithmic checklists based off of what people say work and don't work, except for like, you know, the foundation. And I think the biggest reason is because of AI content, but not the quality of AI content, actually the quantity of AI content. And before where it was like the internet was filled with You know, people could pump out dozens of pieces of content a day, you know, by offshoring it to, you know, the Philippines. 00:31:13.86 evergrowmarketing Now it's thousands of pieces of content a day. And so it's I don't think it's necessarily the all the time. The the quality is more of it is like the quantity, um especially it kind of reminds me of that. 00:31:21.85 Jon _Originality_ai_ Yep. 00:31:27.11 evergrowmarketing um That one thing was like, if you have infinite monkeys randomly typing on a typewriter, eventually one of them will reproduce the perfect Shakespearean novel. 00:31:30.13 Jon _Originality_ai_ The produce. 00:31:36.34 Jon _Originality_ai_ Yep. 00:31:37.02 evergrowmarketing And that's kind of what AI content is to me. 00:31:37.59 Jon _Originality_ai_ Yep. 00:31:39.30 evergrowmarketing So like, it just kind of feels like Google is trying to get ahead of that. 00:31:43.32 Jon _Originality_ai_ ah Agreed. but i Agreed. I think, and and sort of by definition, the content has to be, and you can you know, providing unique data sets and stuff to, to you know, make it make it useful or provide some a level of information gain. But kind of by definition, it's a derivative work of everything that has come before it. And when Google has a push for information gain, it struggles to to kind of achieve that. does it Does it work? Does it rank? Sure. And so I think some some of the then i'll yeah I i'll jump into like understand where the negativity around AI detectors come from. um We wish that they, everyone sort of wants to start using them the same way that plagiarism checkers get used, where there's this sort of like binary classification with traceability proof of plagiarism occurred, it didn't occur. And this sort of combination of 00:32:38.62 Jon _Originality_ai_ this new way of needing to check content that introduces friction into a process that people don't love is is real. And so yeah, AI detectors, including our own, get get it wrong some of the time um in sort of all data sets that have been tested were sort of in the two to 3% range for false positives. That's high when there's consequence attached to to um to a false positive occurring. 00:33:08.29 evergrowmarketing Yeah, so that that was one of my notes too was was like, what are your thoughts and reputational damage from content creators, even like journalists who have a false positive on a piece that they write? 00:33:21.45 Jon _Originality_ai_ Yeah, so we we don't recommend our tool being used within academia for that reason where a score a score alone is should not be used as ah for for for consequential decisions. We offer a free Chrome extension that so if you have a Google document, our free Chrome extension will take all the metadata behind the creation of that document. 00:33:44.11 Jon _Originality_ai_ and be able to reproduce a character by character sort of writing history of that document. And so the AI detection score along with, you know, what we're trying to sort of sit in the seat of the copy editor. They get this piece of text and then they need to make sure that it clears their bar for a quality piece of content. Pleasure is in free, write readability range, spelling and grammar mistakes clear. Was it written by AI or not? 00:34:12.36 Jon _Originality_ai_ And so the AI score is the first tell in that and that investigation of was it written by AI or not? If it's clear, then great. If it's you know like the original, then then it's got it right. 00:34:24.01 Jon _Originality_ai_ and And if all other things check out, then it can move on. If it's like the AI, then we have a ah free Chrome extension where they can go in and watch. basically It's almost like watching the right or right. 00:34:35.44 Jon _Originality_ai_ And so it's this sort of like secondary check to ensure that the content was created by a human. Could that get faked? 00:34:40.42 evergrowmarketing So it's like if you just see the text being just like generated, then it's like, okay, well, no one can write that fast. 00:34:46.80 Jon _Originality_ai_ so So what you see, that it got created in two minutes, and a and within that writing, a thousand words got pasted in. And then a few little edits got done to it, and the and you know the formatting got changed away from the default chat GPT formatting, and it got 100% AI probability score. Pretty hard for someone to convince me that if they were told to write the entire document in the Google document from the start, 00:35:16.66 Jon _Originality_ai_ And that's what you see. 00:35:19.97 Jon _Originality_ai_ I'm going to believe the tool, not not the writer in that situation. 00:35:22.42 evergrowmarketing I wonder if there's a prompt that says like, like at a Chrome extension that you can just plug it into like, Google like Word doc and just be like, write this slower. 00:35:32.50 Jon _Originality_ai_ There is. So like everything that we have built, there's been like the like the opposite, like the the cat and mouse game has has been built. 00:35:39.55 evergrowmarketing That's so crazy. Like, like if you're that smart, just do the work. Like if, if you're going to put that much work into trying to get around something, just do the work. 00:35:45.78 Jon _Originality_ai_ Yeah. 00:35:50.56 evergrowmarketing It's like people end up doing, I feel like people end up doing more work with like their prompting science experiments than if they just wrote what they wanted to write. 00:36:00.77 Cody This is what I just put in that YouTube comment we just got a while back, right? Where somebody was talking about rank and rank and their perspectives on it. 00:36:06.45 evergrowmarketing Oh yeah. 00:36:08.53 Cody And I was like, look. Some people are shady just because they like being shady. 00:36:16.04 Cody There's. 00:36:18.58 Cody There are people who. um Don't go legitimate, just in business in general with whatever it is that they're doing, right? um they like doing shady stuff. um And I'm not saying that this is necessarily shady, right? It's just in general, like, as opposed to doing the quote hard work and doing things the, you know, societally approved legitimate way. They don't do that. And they do it because um for whatever reasons, they don't think that they can do the legitimate way. There's a whole path. But you you get down to the point where some people just 00:36:56.86 Cody don't want to, and they want to do this sort of thing. And that was actually, I'd started writing down notes here, because i John, I wanted your thoughts on this too, knowing from the background that you came from. And now, I mean, you you own a company that determines in people's eyes, right? I mean, you mentioned false positives, but in their eyes, right? And I'm sure in the the group that Jake was mentioning too, they see you as as the guy who's capable of saying this is, 00:37:24.55 Cody fake this is illegitimate whatever words you want to use and then this is not this is raw this is original this is real um but you also you came from this other background first of building these sites making portfolios networks you know um arbitrage that whole thing and i don't i don't know what the question is but i I think I just want to point out and call out that this is kind of heavy on my mind lately because I just had a and the talk in one of my classes recently about content farms, right? um Just content mills and everybody in academia is convinced that these are politically motivated and politically driven. Right. And I just told them to stop. I was like, you guys are so far from the reality of what the world is actually like a lot of these people 00:38:22.10 Cody are just from lower income countries and just trying to move up. And that's it. 00:38:25.96 Jon _Originality_ai_ Yeah. 00:38:27.12 Cody There's no other motivations. 00:38:28.58 Jon _Originality_ai_ Right. 00:38:29.48 Cody There's not an ethical consideration. There's not a moral compass. It's purely, I'm trying to make money on the internet where there's a lot of money. And I will do whatever it is without consideration for the fact that, yeah, this will get indexed in Google. 00:38:42.16 Cody Yeah, there's pro humans and all those sorts of things. um Have you noticed yourself having some sort of like, mental progress or or, or ethical progression as you've gone from, you know, portfolio owner to. 00:39:00.94 evergrowmarketing basically a content mill to like a content mill attacker. 00:39:04.13 Cody Yeah. 00:39:04.71 Jon _Originality_ai_ We were at a content mill all the time. 00:39:05.78 evergrowmarketing you're 00:39:07.48 Jon _Originality_ai_ um um yeah and So I think i think it's definitely been so today there's a couple comments to that. And I think you're 100% right. So I think when i when I was first writing content, I'm like, hey, fun game. like Instead of playing a video game online, I'll i'll do this. 00:39:27.73 Jon _Originality_ai_ And it was like 95% of my focus was on the algorithm and 5% of my focus was on the the the reader. I'd say we have a massive content strategy running at originality.ai, getting million visitors a month organically, pushing incredibly hard on that on that effort and with with success. And I'd say we're 95% focused on on The audience and 5% focused on on the algorithm. And so I think that and that's that sort of inversion hasn't been a stepwise it's been like pretty gradual um over over that period of time so 00:40:03.97 Jon _Originality_ai_ if I were to see some of my sites from 2012, yeah, I'd be embarrassed by them, um as I think. And I'd be, depends on, probably not embarrassed to show it to you guys, like anyone and else that's been on this sort of entrepreneurial journey and experienced that that progression, I wouldn't be embarrassed because they they would relate, but to somebody that hadn't, to the that room of professors that have sort of had that sort of like silver cocoon, I would be embarrassed to them because they would they would judge it harsher than and I think someone that, 00:40:32.69 Jon _Originality_ai_ that has, and I think so there's people that are starting that journey um now, and I think especially people from from um you know low-income low-in income environments our are willing to sacrifice a lot um to to try and move up. And so in terms of what we've seen from our users, it's been really been a few fascinating stories. um I hire people, anybody that says they can beat originality um that could bypass us, they've got a job. um So we hired anyone that says that they can beat us, hire them to to see how they can how they can beat us. um 00:41:12.53 Jon _Originality_ai_ None of them have lasted because no one can consistently. um no No one can with a decent quality piece of content. We changed models recently um and then solicited feedback. one One of the people that was providing the provide some negative feedback to say, hey, I write human. This is now detecting me. I review all the all those comments to try and sort of understand and sift through. like I recognize this person. It was one of the people that I knew that I had hired to use AI. They didn't know that because I use it on a different profile. um I had hired them to beat originality. 00:41:50.74 Jon _Originality_ai_ Using AI so I mean they I knew they were using AI and then they were providing an incredibly negative review about how we were not capable of detecting human content But the reality was that I knew who that individual was I knew that they were using AI and they were just no longer able to bypass Bypass us with the newest the newest model. 00:42:08.17 Jon _Originality_ai_ And so we have this Yeah yeah So if we 00:42:08.42 Cody Wait, let's, let's zone in for a minute. 00:42:10.46 evergrowmarketing Yeah. 00:42:10.61 Cody I want to understand the the politics or the logistics of this. you You hired them with a different profile so that they wouldn't know it was you. 00:42:19.72 Jon _Originality_ai_ Correct. So we have a red team and a blue team. 00:42:20.92 Cody Is it like an undercover boss thing? 00:42:23.22 Jon _Originality_ai_ Kind of. It's like a stealth shopper. 00:42:27.03 Cody ah Okay. 00:42:28.00 Jon _Originality_ai_ So we have a red team and a blue team trying to, there's always a cat and mouse game trying to beat our detection team. And so we have different aspects to our red team. 00:42:39.72 Jon _Originality_ai_ One of those aspects is anybody that is out there sharing how to bypass us or attempting to bypass us, we are eager to 00:42:47.97 evergrowmarketing Just third parties, just random people trying to figure out how to bypass you. 00:42:50.41 Jon _Originality_ai_ Yeah, yeah, yeah, exactly. 00:42:51.59 evergrowmarketing Okay, I got some people in mind. 00:42:55.30 Jon _Originality_ai_ right Random's online trying to bypass detection. there's There's lots of them. um So yes, we will go out and hire them. And then they don't know who they don't they don't know that we are the that company. 00:43:07.18 evergrowmarketing Originality or? 00:43:08.19 Jon _Originality_ai_ Yeah. 00:43:08.81 evergrowmarketing Okay. 00:43:10.03 Jon _Originality_ai_ And then they, 00:43:10.11 evergrowmarketing They think they they they think you're like ah like just someone who's hiring like a content writer or something. 00:43:14.92 Jon _Originality_ai_ correct. 00:43:15.45 Cody So you're John and they hear from Sean and Sean's like, hey, listen up, buddy. 00:43:15.67 evergrowmarketing Okay. 00:43:21.75 Cody I got a job for you. Like, is that. 00:43:23.51 Jon _Originality_ai_ Well, they'll be hired by one of my other companies. 00:43:27.58 Cody Oh, OK, OK, OK. 00:43:28.87 Jon _Originality_ai_ So like as a contractor on Upwork, I'm hiring a writer and then I'm in the job posting. 00:43:29.34 evergrowmarketing as a contractor. Okay. 00:43:35.34 Jon _Originality_ai_ I need someone that can bypass originality. And then it's it is and it's super fun, actually. 00:43:37.91 Cody This is fun, man. You get thrills from this. Do you? I'd be like, what else do you, why do you need to ski if this is what you're doing every day? 00:43:41.82 Jon _Originality_ai_ like 00:43:46.16 Jon _Originality_ai_ the 00:43:46.54 Cody Like. 00:43:46.80 Jon _Originality_ai_ the red the like I had no idea I'd have as much fun as I'm having with the red team, blue team combination. um It's it's a actually really kind of fascinating aspect to this and this sort of cat and mouse game. 00:43:57.21 Jon _Originality_ai_ I've never had a business where it's like every morning I'm waking up and it's like, all right, who's going to punch me first here? 00:44:02.30 evergrowmarketing There's like a scoreboard. that's Yeah. 00:44:04.22 Jon _Originality_ai_ Yeah. 00:44:06.67 evergrowmarketing That's so like, okay. 00:44:06.78 Cody That's fun. That's cool. 00:44:08.28 evergrowmarketing So then going back to the story, so they were like, yeah, I can, I can bypass this and they just weren't able to. 00:44:16.41 Jon _Originality_ai_ So they were able to on 2.0.0. So one of our our older models, um they they had a method that still didn't produce great content, but by producing enough grammatical mistakes, and it still works on a lot of other detectors, by producing enough just enough grammatical mistakes that it reads well, but the detector would be like, no, there's no way an AI produced that. um That was his method of sort of using prompts to say, introduce X number of grammatical mistakes into this piece of content. 00:44:42.50 Jon _Originality_ai_ was basically the the short version on how the prompt worked to to bypass detection. 00:44:47.00 evergrowmarketing Yeah, I've seen things like use, like prompts, like use passive voice, add in like salience and all this and it's, yeah. 00:44:53.12 Jon _Originality_ai_ Yeah. There's no prompting strategy with the exception of producing an outrageous level of grammatical mistakes that will result in AI being capable of bypassing detection. 00:45:08.38 evergrowmarketing So how, how does it, how does per ah originality like detect AI then? Like what does it use as a, as a metric? 00:45:17.69 Jon _Originality_ai_ Yeah, so so the the unsettling comment about this, and I mean, this is the same sort of unsettling comment that that and sort of the makers of any other AI system have, or whether it be Google or chat GPT is like, how does it work? We don't know. What's it looking for? I don't know. what did it Why did it think this piece of content was AI? I'm not sure. um Which is which is a sort of an unsettling comment, and I understand it's not an ideal answer. 00:45:45.54 Jon _Originality_ai_ But the way it works, the way is it so it's called a classifier. So it's an ai AI that has been trained on millions of records of human content, millions of records of AI content. And then it's done what AI is really good at and humans suck at, which is identifying patterns between them and being capable of saying, OK, based on patterns that it's recognized that we don't understand, 00:46:12.55 Jon _Originality_ai_ i have I believe this is AI, or I believe this content was created by a human, and I can provide a confidence score. And then we take that model, so we train, there's um like 50 models, I'd say, with five of them being released. Take that model, and then we have benchmark data sets that we then test against to confirm the efficacy. And then we have third, and most of those benchmark data sets are third-party data sets, 00:46:40.42 Jon _Originality_ai_ And then we have third parties tested as well to evaluate ah evaluate the efficacy of it. 00:46:48.48 evergrowmarketing So as AI content gets better, though, and you don't really know how it works, how does it continually get better and detect AI that because like, you if you train it off of like, ah like, 00:47:04.09 evergrowmarketing Thousands of pieces of AI content versus just thousands of pieces of human written content. It's easy to provide a control for the human written content. Like you can just be like, did you write that? And then like, or like watch someone write something and like know I could produce human content versus like, um, this all this old AI content, this is easy to tell it's AI because it sucks. 00:47:24.21 evergrowmarketing But then like, as AI gets better, how does the model get better? 00:47:24.50 Jon _Originality_ai_ Yep. 00:47:28.78 Jon _Originality_ai_ Yeah, so what we've what so we have to we every time a new model comes out, so like um four outcomes out from from four. Build a a data set off of it. So use the API bunch of different settings, bunch of different prompts and create a challenging data set of what 4.0 can produce that attempts to bypass detection, evaluate our efficacy. If we're not where we want to be, which is sort of like 98, 99% accurate at calling AIAI, then we go and build a much larger dataset to train off of and then close that close that gap. 00:48:07.86 evergrowmarketing So you train it off of content that you produce off of new iterations of these language models. 00:48:15.60 Jon _Originality_ai_ Exactly. 00:48:16.77 evergrowmarketing So then do you have to do it for like each iteration of a like of different language models? Like Chad GPT has theirs and like maybe like perplexity has theirs. I don't know if they're built off the same core language model or what. 00:48:28.67 Jon _Originality_ai_ Yeah, it's perplexical tie into different, like a llama, um, but like Claude, um, llama, grok. So, so the answer is potentially what what's been really interesting to see is like, when we first built a edit action, I'm like, well, this will be fun until GBT four comes out. And then, and then. 00:48:47.70 Jon _Originality_ai_ It won't work anymore. And like, certainly if it's not you before we like, you know, GPT 10 we're not gonna be able to detect it. um What we've seen is that there's been a plateauing of the large language models ability to create diverse content and our accuracy has actually been coming closer and closer on every iteration as opposed to further away. And so when like we were trained on 3.5, four came out, we had a significant significant drop until we retrained and closed that gap. Each new iteration has been not having a significant drop in in capabilities. And every new company that comes out, their model is not wildly undetectable the the way that some it was a little bit earlier on. And so kind of what we're seeing is 00:49:37.69 Jon _Originality_ai_ The same, it's the same data sets. So the common crawl, the common web being trained on the same sort of transformer technology to build these ah LMs on the same hardware. And the resulting language model has a lot of similarities in the type of content that it produces, whether it's cloud or Gemini or llama or GPT-4. 00:50:03.08 Jon _Originality_ai_ And so the answer is yes, we would need to train on every new one that comes out. We test on every new one that comes out and then we train whenever our accuracy drops, but we have not needed to close that gap on training with some of these newest, some of the newest models. 00:50:18.70 evergrowmarketing It seems like they like there's a what's that or that like AI content or LLM seem to be like hallucinating more because they are like the content they're training themselves off of is just like content that that they already wrote. 00:50:35.78 Jon _Originality_ai_ Yeah, sort of the snake snake like that's eating its own tail sort of analogy around like, can can they train on synthetic data or does it just start to go go crazy? Yeah, it's it's interesting. we're We're working with some clients actually that are some of the most recognizable names in the space who are trying to address that problem. i don't I don't know if they'll be able to. Yeah. 00:51:01.72 evergrowmarketing Do you, Cody, you're writing notes. Do you want to. 00:51:03.68 Cody I'm writing so many notes. I'm like the fastest no writer ever seen. 00:51:07.50 evergrowmarketing Do you want more time right now? 00:51:08.22 Cody I got one. 00:51:08.50 evergrowmarketing So do you want to start asking? 00:51:11.98 Cody I think I got two that are that are pretty heavy. Let's split these up. I'll do this first one. um 00:51:18.90 Cody I want to talk about the critics a little bit, a little bit more. um Because we're. We're we're in neither crowd here, um mostly because we haven't used your tool. right um And we i mean we we did ah some relatively even minimal AI content writing exploration. And then at this point, our current stance is really, look, use chat GPT as you would like. Don't copy and paste, but use it for inspiration and and pull things from it. um 00:51:55.53 Cody similar to the, the rubber duck with the developer concept, right? If you need something to talk to and have it talk back to you, that way you can just do the mental processing. It's a good way to do it. Um, but no, don't copy and paste. and So that's kind of our stance. Um, so when I think about people who might say, I tried it and it sucks, right? If that's the the extent of their, their monkey brains, um, 00:52:22.65 Cody My assumption is that it reminds me of that, that Louis C.K. um Comment. Wait, I think he's talking to. um He's on one of the night shows, but he says like everything is amazing and nobody's happy because he's in the the airplane and then he he sends a ah text and it doesn't go through and he's like, this is awful. 00:52:42.17 Cody This is terrible. And he's like, you're in space. Give it a second. Right. 00:52:46.81 Jon _Originality_ai_ Yeah. 00:52:47.00 Cody I imagine people probably have one false positive or something and they're like terrible, right? Like that's that's what I would guess. I don't know, but that's just what I'm assuming. um What are people expecting, right? what are What are they assuming when they sign up for your product? And if they are writing into something like that, just what is the disconnect or is there just a lapse between expectations and reality or or what what do you think is going on? 00:53:15.75 Jon _Originality_ai_ So I think we have a very loud, so so we wrestle with those all the time. And so I'll share some metrics that we look at. the the metric and There's a metric in SAS around net revenue retention. So for every dollar that we earned last month, what was that what was that same and for the customers? Let's say 100 customers, and we made $100 from those customers. Some will churn, some will grow. how much How much did we make this month from that same crew if we hadn't had zero new customers, but some churn, some increase or spend. um So with that, net revenue retention is sort of 00:53:50.33 Jon _Originality_ai_ 90 to 100 is like good, especially in like business to consumer. e We're kind of like a light B2B, kind of B2C-ish, mostly B2B. You know, over 100 is great. I mean, so we've seen, if you're like a quick, if you're churn and burn kind of SaaS, you're going to have a net revenue retention that's really crap, like 70, 80%. Our net revenue retention is, puts us We're but above 100% showing that the customers that are using us are gaining an increasing amount of value by by using us. And where we have customers that have been with us forever. When we look at our ah or so churn rate, it drops off to great B2B metrics. Early churn rates are very, very high. And so why why are we why are some why do some people hate us? 00:54:42.31 Jon _Originality_ai_ writers hate us because this is a ah thing that they didn't have to do before, and they're fearful for AI taking their job away, and they are, I think, my my view, because I like, we hired writers, I work with writers, I like writing, I thought we were building a tool, and I think we are building a tool that protects writers, but they view like this sort of AI were the tool at saying your content was human or your content was AI two to three percent of the time. um That is hurting them but if it wasn't for AI detection then anybody that had access to the internet could be creating content at a pretty decent quality that would certainly outstrip my capability of writing and so I think we in to some ways we're defending writers but I understand when we 00:55:37.69 Jon _Originality_ai_ Somebody pours their heart into a piece of work. It gets called AI when it wasn't. that is a And that is their profession, that they are fearful of being taken away by AI. That's a very emotion-ridden experience that we are to blame for. And I understand why the sort of level of anger comes out. And so I think that's a part of the very loud anti-detection crew. 00:56:06.67 Jon _Originality_ai_ writers that want to be making a If you're a writer and you're making $20 an hour, $30 an hour through your writing efforts, and you can double your productivity by using AI, you want to do that. ah mo like To your point, these writers are but some of them are doing it for the passion, but if you're in that sort of 20, call it 50 an hour kind of range, you're probably a part of your job is to just get the work done so that you make money to support your family. And this tool is making your job. If it wasn't if it wasn't for AI detection, I'd be able to use AI and still get paid that. And that's sort of the the the thought process, but you extend that thought process out a little bit longer and say, without this AI detection, 00:56:53.49 Jon _Originality_ai_ Oh, I'm going to get out competed by people that are willing to do this for way cheaper because there's no defense which behind my ah written work. So I think that's another significant piece of it. And then I think there's others that another sort of loud segment are people that want to use AI for publishing content on Google. And there's sort of a crowd, you know, crowd mentality that can kick in where Like we're the ones saying, you know, we're not Google, but we're the ones saying like, no, I think Google's not cool with youth publishing a thousand pieces of content a day that no human has reviewed. And then it's like, you know, pitchforks and pitchforks and torches to say, no, that's that you're wrong. But look, they said over here that they're fine regardless of how the content was produced. And look how cool and clever my prompt is. I can totally publish a thousand posts today and, and not be hurt by Google. 00:57:50.46 evergrowmarketing So what you got a, you got a mob of lazy people after you. Okay. Good luck catching up. But I, I love that. 00:57:56.04 Jon _Originality_ai_ Right. 00:57:56.36 evergrowmarketing My, the, the fate, my favorite part about your argument there or like the, the, about the people that your haters is you open that up with doesn't change what happens in my Stripe account. 00:58:09.91 Jon _Originality_ai_ Well, it does but my point my point is is isn't that like, 00:58:09.97 evergrowmarketing yeah 00:58:16.15 Jon _Originality_ai_ We're an incredibly low price product that is incredibly easy to churn from. And if and we're we're attached to a value metric around credits being used. And if that credit's being used is going up by on ah on a cohort basis, then we are doing something right to the customers that are paying. Because they could if our customers had the same reaction that the community happened had, our net revenue retention would not be where it is. 00:58:40.94 Jon _Originality_ai_ we would we would have a it's it's a Intentionally, we're were not one year contracts at some inflated fee that doesn't get used. If our tool doesn't get used, if our tool isn't producing value, then we don't get dollars into our Stripe account. 00:58:56.06 Cody I'm going to I'm going to share this with people at the university, because this this entire like how this is covered into writers, especially. um This has been a topic that has come up a lot of times, and it's it's exactly what you said. I mean, you have I think it's a minority of of the writers, because I guess context here, um the Iowa Writers Workshop is is supposedly one of the best writing places in the US, right? um They put it in the I think it's number two. 00:59:26.04 Cody Um, last day checked and they take it very seriously. Um, specifically it's creative writing and that's what scares people the most is because it's not, um, nonfiction, right? I mean, this is people who are writing their souls out and it's the minority that says AI in general is okay. Um, in, in any capacity, because you have the other crowd and I think they're more vocal who are saying, 00:59:50.57 Cody This is Satan, right? like but That's truly their perspective is this this has no place here and with what we're doing and what not. um I wanted to ask very clearly plagiarism versus AI detection, right? 01:00:08.12 Cody Because to these writers, plagiarism being the the verbatim copy and pasting, right? 01:00:15.09 Jon _Originality_ai_ Yep. 01:00:15.32 Cody And of course, the all colleges are concerned about that. um But then the individual writers, regardless of academia, have what you've said, a fear, probably, of of being mislabeled with AI detection. What's the difference? What is the difference between AI detection and plagiarism detection? 01:00:36.28 Jon _Originality_ai_ So I think we've 01:00:42.21 Jon _Originality_ai_ So the the like I'll give it like call like a textbook definition, but like plagiarism is if if work at work already existed and was copied and pasted um or paraphrase plagiarism exists where it's just like synonym swapping out some words and and then that's that's the been the new work, but really it's just the the old work. Or it's copying and pasting a bunch of different paragraphs and that's patchwork plagiarism. Plagiarism is copying other work and and pat trying to pass it off as your own. 01:01:09.57 Jon _Originality_ai_ That gets identified in in our tool by looking at each piece of content using some algorithms to look at what's out on the web and then comparing um the similarity and then identifying those those matches. that's That's plagiarism. AI detection is, was AI involved in the creation of this piece of content? And if you want truly human-generated work, then it needs to be both not generated by AI and not copied from from somebody else to get that specific person's unique original work. And so that' that's the the difference, but also why they are um why they should be viewed, but why why we it made sense for us to put them in the same tool because a copy editor wants to know that the piece of work that they're paying someone for is is is what they were promised, which is their own unique research to their own unique view and their own their own unique output. 01:02:04.50 evergrowmarketing There's that argument that like just because it's produced by ai generate art it's it's AI-generated doesn't mean it's not plagiarism because AI is trained off of other people's work. and so theyre fo Therefore, if you produce AI-generated content without any of your input, like and I mean like without the majority of your input, then in a sense that is plagiarism. 01:02:14.95 Jon _Originality_ai_ Yes. 01:02:25.74 Jon _Originality_ai_ Yeah, I would say that is that like, ah I don't know where I land on that argument, because isn't isn't all work, just a isn't a derivative of work that has has come before? 01:02:40.09 evergrowmarketing Hmm. 01:02:40.80 Jon _Originality_ai_ Like, you know, this, this conversation is not AI generated, but is clearly the off the backs of other knowledge and and learnings that we have we have had. And so are we plagiarizing from the group that we're, we're quoting from? I mean, so I think, I think there's, there's an argument there on whether it's plagiarism or not. And I think that whether the AI output is plagiarism or not, but I think that's like a to be determined by the courts. um And I see, I see both arguments and we don't, we don't, we don't call AI plagiarism. 01:03:13.81 Jon _Originality_ai_ We call it AI generated. 01:03:17.38 Cody I love that. I love. I just wanted to show like everything is a copy of a copy, you know, like that because it I don't know. 01:03:22.08 Jon _Originality_ai_ Yeah. 01:03:24.94 Cody it It sounds like brosudo signs a little bit, but I feel it. I feel it deep in my soul of like when I go to write something anymore now, right? um It feels like it's already been written somewhere. 01:03:37.58 Cody And if I'm trying to say something, there's probably already but already somebody out there who has said it better. So to truly write something unique and original. um takes extra time because there's a unique and original in general, right? As in the closest so you can get as has never existed. But then there's that's the harder one. Then there's also just the what's the hole in the Internet? What can I write that isn't showing up already? Something that um isn't going to take us especially 01:04:11.10 Cody um And you think about content rating for the purpose of ranking, something that isn't going to take a lot of backlinks in order for me to show up. What's a what's a gap that I can find that's a low competition, that sort of thing. And that's I find that one a lot easier. I have. I have one more question, but it's kind of heavy, so I don't know, Jake, if you have other ones. 01:04:31.00 Cody um 01:04:32.45 evergrowmarketing Um, yeah, I mean, I got one that I kind of wanted to to ask too. So when I was looking at kind of how AI checkers work, um, there, this says, this is from search engine land, that there are two things that AI checkers really look for. There's perplexity and burstiness. So perplexity is basically referring to just their unpredictability in writing. Um, 01:04:56.54 evergrowmarketing I'm assuming that's more like, you're not using this like the same adjectives or even an adjective for every noun you have. And then burstiness, which would be like capturing the rhythm, ah the variations of different different sentences. So I'd be willing to bet, excuse me, choked on my own saliva there. ah I'd be willing to bet that AI could, AI would 100% tell me and Cody are human writers. 01:05:23.51 evergrowmarketing Because the way that Cody and I write, we have a very burstiness, um, form to our writing, like we'll do one sentence or one word sentences just randomly. Um, like it'll be like content, like a one sentence. And then it'd be like gap the word stop at a period. And then like the next set, it's like, there's very, like, there's a lot of personality to that. Um, 01:05:49.47 evergrowmarketing And so I just wanted to ask you, man, maybe this is just a yes or no question, not really a good podcast host question, but, um, you'd mentioned that you didn't really, your models are trained off of AI tools. So like, really you don't know what they're looking for, but would you agree or disagree with perplexity and burstiness kind of being like the core foundational elements of AI versus human? 01:06:16.97 Jon _Originality_ai_ um There's a lot, so disagree, but I can give some context. So that article is wrong, but I understand why it's wrong. 01:06:27.21 evergrowmarketing Was it written by AI? 01:06:27.78 Jon _Originality_ai_ The, hallucinated, just made some crap out of it. 01:06:36.32 Jon _Originality_ai_ It was a one of the other early tools that went viral. kind of relaunched just before chat should be T launched. And then another tool launched after they were focused around, um, call it this like bag of words approach where they take this sort of algorithmic approach to say, what are a bunch of markers right now with AI content? Um, perplexity burstiness were markers of, of AI content readability scores. So most AI content ends up in a fairly closely normally distributed 01:07:10.63 Jon _Originality_ai_ curve around specific readability scores and so if you can the sort of those a detectors which have all sort of now moved to the the the classifier training model because the bag of words approached. 01:07:26.07 Jon _Originality_ai_ didn't didn't scale well with the newest models, um would look for trying try and put a whole bunch of these um indicators together, where it's like if the readability range is within here, if the burstiness is here, if the perplexity is in this range, therefore it's probably AI content. And so the answer the answer is no, that article is wrong. um But some of the earliest detectors did look for for bag of words or sort of like to make a rule based approach on why AI was being called AI. 01:07:59.21 evergrowmarketing This was written on September 24th of this year. This is wild. So like you're saying that that's outdated. Like that's like. 01:08:06.31 Jon _Originality_ai_ So so that's that's like a 20, that's knowledge that persists from early 2023. 01:08:13.23 evergrowmarketing I was like, this is like three weeks ago. That's wild. 01:08:15.30 Jon _Originality_ai_ Yeah. Yeah. And like, you know, I've been, I've been shocked at, you know, there's, there's, 01:08:17.35 evergrowmarketing Man. 01:08:21.41 Jon _Originality_ai_ You know, it's a probably a broader topic, but I've been shocked in this world around like some of the misconceptions that have been generated that have there was a study done. um That looked at our factors biased against non native English writers. This has been this rumor that has like this really crappy Stanford study that looked at 97 articles to try and assess a 2% false positive rate. yeah And they took it from a student forum um that post-dated GPT-3 but predated chat GPT and then therefore called it human, but they didn't control for would those students have been using AI 01:09:05.42 Jon _Originality_ai_ and then compared university entrance essays to grade eight student essays. And anyway, this is a terrible study, um but because it was published by Stanford or Stanford, it it has, and it's this beautiful fit around AI being biased against underrepresented um people that it has just sort of, it was wildfire in terms of how how persistent 01:09:13.26 evergrowmarketing yeah 01:09:35.24 Jon _Originality_ai_ this, and we've run our own studies looking at control data sets, yeah you know, tens of thousands and, and there is, you know, we're still wrong, you know, three, 4% of the time, um, but no more on, on non-native English speakers. Anyways, it's, it's interesting. This world moves so fast that certain, um, things that get embedded into the world take a long time to, of the, into the knowledge, take a long time to get corrected. 01:10:04.04 evergrowmarketing It's crazy the the stupid stuff that you can spout and be taken for as fact when you have credentials behind your name and like and like nobody will challenge it. 01:10:14.93 evergrowmarketing And then even if you do, you're going to get buried under it anyways and no one's going to pay attention to you. 01:10:15.24 Jon _Originality_ai_ Yeah. 01:10:19.22 evergrowmarketing like Who's John Gillum? 01:10:21.12 Jon _Originality_ai_ Yeah. Yeah. Like I've, I've like written like, be careful calling Stanford study crap. I'm like, but like, look at it it. It's like, you can't tell me that 97 cent, like how it doesn't take any level of thinking to look at 97 samples to try and identify a two, 3% false positive range and, and not call that study crap. 01:10:33.44 Cody Heh heh heh 01:10:41.35 evergrowmarketing Yeah. 01:10:42.35 Cody Well, two, 3% at a hundred is two or three literally. Right. Is that. 01:10:47.64 Jon _Originality_ai_ yeah right 01:10:49.37 Cody Oh, man. 01:10:51.46 evergrowmarketing ah Yeah. I had, uh, uh, I can't remember what I was going to, I had something, I'll remember it later, but I was looking at some of your other stuff too. Cause you're not just obviously an AI detector or a plagiarism detector. 01:11:03.63 evergrowmarketing Like you have other free tools on here, on your, on, on your site. 01:11:08.36 Jon _Originality_ai_ Yeah. 01:11:08.45 evergrowmarketing And I wanted to call it a couple of them out cause I'm like, Okay. This could be useful, like capitalization tool. 01:11:14.02 Jon _Originality_ai_ All right. 01:11:14.32 evergrowmarketing Okay. That's semi-useful. Right. I have to go all the way back through. Like, if I want to do like sentence case or cap cap case, eulogy generator, who asked for that? Who was like, who won? 01:11:25.95 evergrowmarketing Who won the originality team was like, you know what? Uh, I don't want to write this, but I. 01:11:32.09 Cody I don't know if I want to know the story though because how does that come about? 01:11:33.09 evergrowmarketing yeah 01:11:37.17 Jon _Originality_ai_ I probably, probably not that cleverly. Like, uh, I think it would have been a, you know, a Q a content, the content strategy team works hard, um, identified, um, you know, through whatever means keyword tool competitors looked at what, what was, you know, anything that we can rank for that is related to writing, um, and people that want to generate something with, with AI. 01:12:00.44 Jon _Originality_ai_ Okay. 01:12:00.85 evergrowmarketing So the, the free, the free tools you're creating are like, there's an SEO strategy to it then. 01:12:00.98 Jon _Originality_ai_ Not, not a. 01:12:05.99 evergrowmarketing So you're like, cause there's a ton of them on there and I'm like half of these, I feel like are absolutely useless. 01:12:06.48 Jon _Originality_ai_ Yeah. 01:12:11.64 evergrowmarketing Like, uh, but you, I think you're well aware that like the strike through generator, why would I just type it into word doc and strike through it? 01:12:11.79 Jon _Originality_ai_ Yeah. 01:12:18.16 evergrowmarketing Like, yeah. 01:12:19.29 Jon _Originality_ai_ You don't have word, don't know Google Docs, for sure. And you want to just text, you use it to text. 01:12:22.43 evergrowmarketing yeah 01:12:23.85 Jon _Originality_ai_ So yeah, I'd say we're in the like spaghetti against the wall stage um with this with this part of the content strategy. And then our plan is to, like we've had some that have done quite well, that have been received well. 01:12:39.29 Jon _Originality_ai_ I mean, that we've cast the net. 01:12:40.29 evergrowmarketing When you explain it this way though, it like, it makes sense, like a hundred percent. When I was looking at it on the tools, I'm like, yeah, you just sat there and like created these like semi useless tools. 01:12:45.54 Cody That's brilliant. 01:12:49.20 evergrowmarketing But I'm like, no, no, no. Like there's demand for it, which is wild. 01:12:52.56 Cody ah It's, it's almost the equivalent of doing a keyword for campaign for the, you're like, Hey, if it's cheap and you know, and and it's, it's close enough, then we'll get the, the traffic and, you know, as long as the hosting is cheap enough. 01:12:53.63 Jon _Originality_ai_ There's demand for all of them. 01:13:05.73 Jon _Originality_ai_ Yeah. Yeah. And then we're going to turn some of them in like some. We have a plan around some of them that are more closely connected to, like again, like if you take a copy editor and look at all the tasks that they do, there's a lot of, we're trying to build tools that fulfill all that. Like we have a text compare um free tool, which we'll be integrating into the plagiarism checker. Cause that's been you know well received and then we can see a fit. So anyways, we it's like kind of the spaghetti against the wall. We'll probably take the craft that has fallen on the floor off the site and then improve on the stuff that stuck 01:13:41.12 evergrowmarketing I want to meet the people who are Googling eulogy generator. 01:13:45.51 Jon _Originality_ai_ we 01:13:45.63 evergrowmarketing Oh, I don't know. They're still, who are you writing this for? Why are you writing this? 01:13:49.67 Jon _Originality_ai_ yeah don't i don't I wonder in today's. 01:13:51.18 evergrowmarketing I, Mike, Mike, so I gave my, I gave a eulogy at my grandma's funeral last December and, uh, it was like an open thing. So like you could come up and just like, say a few words and like, that was your eulogy. 01:14:03.01 evergrowmarketing And I wrote mine like. I didn't know we were going to do this open mic thing. So like I wrote mine in my seat, just on my notepad on my phone. And it was very heartfelt and I read it and I made people ball. 01:14:16.53 evergrowmarketing And I'm like, I don't think AI could do that, but also like, why would I even want to stand up if I'm just going to have AI generate it for me? Like that's just wanting to hear myself talk at my grandma's funeral. 01:14:28.29 Jon _Originality_ai_ so so So you guys are writers, right? like like both of you like our I would rather communicate in spreadsheets. 01:14:31.94 evergrowmarketing Yeah. Self-proclaimed, sure. 01:14:34.12 Cody but This has all gone in the direction that I originally wanted it to. So I have the question. So that's what I was going to ask you, John, was do you write? Are you a writer? 01:14:45.56 Jon _Originality_ai_ So yes, I have to write. like I write for work all the time. If my like high school English teacher knew the volume of kind of content that I was responsible for, they would be terrified. 01:14:57.31 Jon _Originality_ai_ the that I write because I have to, but I use, I would rather communicate in spreadsheet format. I'm a mechanical engineer by by heart. and And so I do use, I would 100% use ChachiBT if I were to do a eulogy, because I would ask them, I would i would do some version of what you're doing to like this like the pour the heart out phase. 01:15:20.26 Jon _Originality_ai_ And then I would say, make some variations of this. 01:15:22.73 evergrowmarketing So you searched this, this was, I bet you went to your content team and you're like, we need this tool right now. 01:15:23.03 Jon _Originality_ai_ And then 01:15:25.98 Jon _Originality_ai_ that's That's why we created it. I need this. i oh But yeah, I i would 100% do it. like i and like In some and any communication where I'm sending an email, I'm usually dumping my own thoughts down and then asking it to to give me a ah version of it. um i feel I feel great about that. I'm not discing disclosing that I used it, but if you ran my content through originality, you would show up as AI generated. 01:15:57.59 evergrowmarketing But I, I get that for sure. 01:15:57.84 Cody do i think man i think that uh 01:15:59.88 evergrowmarketing Sometimes I don't like creative juices aren't flowing for me, but I have, I've developed systems pre that predates AI. Well, chat GPT AI, uh, that predates that into, that helps me write the content that I want. 01:16:23.10 Cody in business, there's all the different ways to do business, right? um But if we're assuming that you start with what Jake and I especially like, which is no venture capital. but I mean, bootstrap is an ironic word, but as bootstrapped as you can be, use what you have, build up something first, and then use that to go to the next steps. And during those next steps, that's when you can make more conscious, um ethical or moral decisions, I guess, not to say that you shouldn't do that at the beginning. You can, if you should, but you might not have that luxury, right? Early on, early on, it's just, I need money. So once I have some money and some connections, once you have everything that comes with building business one, you get the power, I guess, and the control to make some of those other decisions. Um, which I wanted to ask. 01:17:20.77 Cody You know, I get that you're a spreadsheet guy, but and you're not a writer. But do you think about this philosophically? Do you do you spend the time and consider white what these things do and how how they impact beyond the business perspective? Because there's ah there's a cultural and societal thing here too, right? When we were talking about Language complexity. 01:17:50.41 Cody Jake used the words. What was it? Burstiness and blanking on it. 01:17:53.00 evergrowmarketing ah burstiness, and what was it? 01:17:53.75 Jon _Originality_ai_ Perplexity. Perplexity. 01:17:55.90 evergrowmarketing You know this one, right? 01:17:57.37 Jon _Originality_ai_ Perplexity. 01:17:58.74 evergrowmarketing Was it perplexity? 01:18:00.05 Jon _Originality_ai_ Yeah, first need some perplexity. 01:18:00.55 Cody Perplexity. Yeah. 01:18:01.50 evergrowmarketing Yeah, perplexity, yeah. i was I didn't want to say it because I'm like, that doesn't sound right because of that perplexity search engine. 01:18:03.02 Cody um 01:18:08.05 Cody Yeah. ah My general feeling about writing, and I've grown to be more into this instead of the original, it's just weird being back in college now because when I'm being fed is the opposite. I like accessibility. I like accessible language, knowing that people might not be reading it in their first language. So as simply as you can convey messages, I think that's the way to go. 01:18:34.02 Cody Now, I don't know, and I guess this is a question too, is more complex languages is using those, and maybe this is but that definition of the bursting as perplexity as using those unusual words in unusual places. Is that unique? Um, and is it good? Because I, I believe that when you do that, you become exclusionary, right? When you start using language that people aren't familiar with. 01:19:05.17 Cody And this is an education thing, right? More quote, educated that it comes across. You're no longer being accessible to the majority of people. You're, you're actually writing to an elitist group. Do you think about this? 01:19:19.35 Jon _Originality_ai_ so I'll answer the first guess the second part first and then and then sort of the like like the fairness, societal component second, because I think it's a long one. 01:19:29.85 Jon _Originality_ai_ and think that the I think highly complex, like ah me, you know, again, going back to engineer, like I'm not, if if I'm looking at a spec for, you know, when I was doing engineering stuff, I want the language that communicates what is necessary as effectively and accurately as possible. And so if that involves complicated words to, in order to do so, then it's my job to understand what those what those words mean. 01:20:00.55 Jon _Originality_ai_ If it's, if it's i don't want I don't want this sort of and instrument build manual for a nuclear reactor to be written at a grade five level. I want to be written at the level that is necessary um to to ensure that it gets done correctly. And so I think it depends on on the audience. I'd say it should always, you know I think it's generally accepted that it should be written at the lowest level possible to communicate. you know It's like the saying, like you know sorry Sorry this letter was so long, I didn't have time to write a shorter one. I think that the simplicity should always be pursued, but with no sacrifice of of effectiveness of communicating your point. I'd say that's my my view on on that piece. 01:20:45.15 Jon _Originality_ai_ the sort of like fairn like you know Do I think about the the fairness, the societal importance of and the impact of generative AI and then of our company in particular. I did not set out to build a societally important business, and it's a little bit grandiose to say that it is, but but there's certainly a a ah need for, you know do I want to live in a world where We cannot tell the difference between a AI content and human content. 01:21:19.59 Jon _Originality_ai_ And the answer to that is no, and we've been focusing heavily on writers. But what I'm looking at, and we did some work for a ah journalist looking at a mushroom book, a foraging book that was written by AI that recommended people to taste it to determine if it was poisonous or not. 01:21:36.14 evergrowmarketing Hmm. 01:21:36.79 Jon _Originality_ai_ Not really a good idea. 01:21:38.02 evergrowmarketing I mean, ah very effective, though. 01:21:40.47 Jon _Originality_ai_ You will know, or at least the person with you will know if it was poisonous or not. um Yeah, so I think, yeah, I think there's tons incredible, incredible benefits. I'm glad generative AI exists. I think I want to live in a world where we can tell the difference between human content and AI generated content. 01:22:00.79 Jon _Originality_ai_ Um, you know, reading baby formula reviews, probably don't want those to be generated. Probably want to would rather deal with the, the the messiness of human language, um, a human writing in and read that review. Um, so yeah, I do think, and then in terms of the fairness within the world of that, we're sort of like been more focusing on between this interaction between, um, writers, editor, agency, client. Um, you know, I think fairness is a really, 01:22:30.29 Jon _Originality_ai_ important thing to exist. And, you know, I kind of asked you guys, like, how would you guys feel if you found out that one of your writers was able to produce the content in a minute of work on chat GPT? And I don't know if they get paid um per hour or per piece of work, but there are kind of a paid per word or per per thousand word basis. And you find out that they were copying and pasting, even if it's great content and serving your needs probably wouldn't feel fair. 01:23:00.60 evergrowmarketing Oh yeah, we switch them to hourly. 01:23:04.27 Jon _Originality_ai_ Right, teach us what you're doing and let's switch and switch them to, switch everyone, yeah. 01:23:07.59 evergrowmarketing Yeah. 01:23:11.39 Cody So there was ah there was one more thing that we talked about in the discovery called, I just want to um bring up as a topic because ah it's kind of about that paying AI written content versus human written content. What are your feelings on it? What's your perspective on Because the the gut check, right? 01:23:33.33 Cody Is that it feels cheap. You're like, you know, if, if this wasn't written by a human and human thought and effort and sweat and tears didn't go into this, that I shouldn't be paying a premium for it. That's how it feels in the gut. 01:23:44.15 Jon _Originality_ai_ Yep. 01:23:44.47 Cody Right. Um, what do you think about it? 01:23:47.90 Jon _Originality_ai_ I think that's totally true. I mean, I think there's a ah balance between like the. The value that you're contributing as ah as a, let's call it a writer, as an agency, that if you're you know even even if even if you're happy with it on your end, if another agency is then selling it for less because it makes sense for them to do so, um you know who gets to count if if you're happy with the generated content, and who gets to capitalize on the value, that the efficiency that's being generated. 01:24:16.39 Jon _Originality_ai_ and i think that If that's not explicit, then it's whoever's grabbing grabbing it first, which is often the writer. So I'd say that's kind of a ah piece to that story. I'd say, you know, I think I'm i'm great with AI content. 01:24:34.89 Jon _Originality_ai_ If there's a human in the loop and you know we have our AI research team, English as second language individuals, create some great research documents and papers and studies, um all of the content will show up as AI because has AI has been used in the in the writing of that that content. We happily publish it um we know it. I believe it produces a little bit more risk for Google because we're publishing a lot of content. 01:25:01.87 Jon _Originality_ai_ And some of it is AI, and I think that is riskier than if that was all human, all no, no AI involved. I'd certainly believe that in the eyes of Google, that's, that's a riskier play that we're doing by having some content be AI. um But, but we're accepting that risk. 01:25:17.37 Jon _Originality_ai_ So anyway, I'd say it's, if if I can, i think I think moving forward, like, you know, how is society going to evolve to deal with this world where we don't know if it's human content or AI content or or struggle to know what certainty. And I think the author behind work is going to become ah a more and more important piece to knowing, do we want to care about the piece of content we're about to read? 01:25:42.10 evergrowmarketing I think my final question is there there's talk about legislation regarding word marking AI, text AI, especially like image generated AI and video video generated AI, and even now that we're getting into YouTube, YouTube specifically asks if AI is used to feature any images or people or events in that video. 01:26:05.10 evergrowmarketing and more platforms are becoming hyper aware and wanting to create transparency between AI generated, in this case, content and images. ah What do you, I don't know how to phrase this question, but like, what do you one, what do you think about word marking AI generated text? Can it be done? What does it look like? And um how far out do you think we are from it? 01:26:32.14 Jon _Originality_ai_ yeah So can it be done 100%? Easy. There's open source libraries that would do it now. um How far out from it? I don't think it's ever going to happen. the the And then like why? 01:26:48.43 Jon _Originality_ai_ the whoever So the language model that created the text would be the one that would need to have the algorithm to understand it. It'd be immediately able to be um overridden by another model. And so with there being open source models out there, um lama mr l um there's no, I don't see a scenario where everyone would get on board with having the same watermarking algorithm apply to all of their content. And it would be, ah ah there'd be a piece of complexity to it that would reduce the quality. And so I just, I don't see anyone getting on board with that. And I also don't see there being sort of the 01:27:31.15 Jon _Originality_ai_ The political appetite for that being where watermarking occurs in in the form of text, I think think a single document of text is very low consequence to society. 01:27:45.16 Jon _Originality_ai_ you know I don't think it's easy to sort of like hold up a declaration of independence and be like, oh, this was pretty high consequence. um But I think like you know in today's today's text creation, I don't think you can have a single document that is going to be societally impactful. 01:28:01.23 Jon _Originality_ai_ as opposed to a body of work, there's other problems that will come from mass produced AI content to move society. um But I think images, video, audio have a far higher societal consequence to a single piece of content. And I think the watermarking sort of political appetite to legislate something is going to land or more in other forms of of content than text first. 01:28:29.32 evergrowmarketing Do you think your business model benefits off of not having a watermark technology? 01:28:39.98 Jon _Originality_ai_ I think, it now you know, if if if my business wasn't needed, would i would I be happy? And I don't know. like I mean, I'd be bummed, right? The financial consequences would suck. 01:28:50.19 evergrowmarketing Well, you still have a eulogy generator. You could just paywall that. Okay. 01:28:54.61 Jon _Originality_ai_ And with the Terminator coming, it'll be fine. 01:28:54.69 Cody Just make it premium. 01:28:57.17 Jon _Originality_ai_ well We've got got lots and lots of death. 01:28:59.33 evergrowmarketing so yeah So I know that I can count on you lobbying against toward marketing. 01:29:08.47 Jon _Originality_ai_ yeah So I don't know what would happen. um So I think if if watermarking were to be enforced, it would be enforced by at a US level, um probably first or a European level first. And then there would be open source coming out of other locations that would be very easy to unwatermark. And then AI detection would still be potentially more needed because water the it would mean that there was a heightened sensitivity around making sure that AI content was identified as AI. 01:29:38.38 Jon _Originality_ai_ yet watermarking would become instantly ineffective at doing so. So I think it's like, got to think a couple steps ahead on that one, but I think it would be beneficial to the business because watermark, the society was, it was so concerned that they enforced watermarking, but then watermarking would fail instantly, and then we would be back to imperfect detection at identifying who was AI or human. 01:30:00.73 evergrowmarketing Yeah, I kind of think I'm with you. I think any private entity in the US that comes out that if if there was some kind of watermarking legislation, I think it would just be on a company by a company basis because they're not going to you can't do that on a person by person basis on whether you use it or not. 01:30:13.95 Jon _Originality_ai_ Yep. 01:30:19.78 Jon _Originality_ai_ And you wouldn't be able to apply it globally with open source out there already. 01:30:24.14 evergrowmarketing Mm hmm. 01:30:26.00 Cody John, I think you just sell it, sell the business and then ski as much as you want before then, before it becomes your problem. 01:30:26.15 Jon _Originality_ai_ Yeah. 01:30:31.71 Jon _Originality_ai_ It's not a bad idea. 01:30:34.34 Cody Yeah. 01:30:36.81 Cody Just call it good. Be good with it. That'd be nice. 01:30:40.28 Jon _Originality_ai_ ah the the red the you know Full circle, but the red team, blue team is too much fun right now. 01:30:44.67 Cody Yeah. 01:30:45.11 evergrowmarketing Yeah, yeah, yeah, for sure. I think it went on our discovery calls. Like I would love to be on the on the red team, right? They detect, right? they They try to, they're trying to, or which one were they doing again? 01:30:55.37 Jon _Originality_ai_ red Red team's the like the offense. they're trying to beat They're trying to beat the detector. 01:30:58.35 evergrowmarketing Okay. 01:30:59.61 Jon _Originality_ai_ So they're trying to like find find ways to... 01:30:59.95 evergrowmarketing That would be, I, yeah, I would, it would be fun to be on the red team. 01:31:03.37 Jon _Originality_ai_ Yeah. 01:31:03.39 Cody Yeah, that's fun. 01:31:04.75 Jon _Originality_ai_ I'm totally useless on the blue team, but I'm enough of a sort of a shit distributor that I have some some effectiveness on the red team. 01:31:12.59 Cody Cool. 01:31:14.05 evergrowmarketing All right, well, that's all I had. Cody, you got anything? 01:31:16.48 Cody No, this was a, this is cooler than I thought it would be originally, which isn't meant to be an insult. I just thought it was like more fascinating than I was prepared for mentally. 01:31:28.35 evergrowmarketing ah John, i what do you, just leave us with your final thoughts specifically to the the haters of AI content detection and originality AI, and then tell us where people can find you. 01:31:43.63 Cody Yeah, plug yourself. 01:31:43.88 Jon _Originality_ai_ So like yeah the so the the haters that are that are like genuinely hurt by a false positive, you know I apologize. and you know I feel that pain. to To the haters that think a detection bullshit doesn't work, we have an open challenge to anyone that wants to a to challenge the tool where we will benefit charity. New data set, everyone we get right, you donate to charity, everyone we get wrong. 01:32:07.55 Jon _Originality_ai_ um we We'll donate to charity. So if anyone you know really believes um with with, you know, beyond just sort of a grow science test of one of one piece of text that that they don't work, then yeah, we have an open open challenge. to Anyone that wants to take us up on that, that'll benefit charity. No one has yet, but we're hoping one day. 01:32:26.32 evergrowmarketing I would love to see that. I'd love to see someone take that on. 01:32:29.31 Jon _Originality_ai_ I would too, it'd be great. 01:32:31.09 Cody Will you stream it on YouTube? 01:32:31.44 Jon _Originality_ai_ So yeah, nope. 01:32:32.45 Cody I would sub I'd watch for sure. 01:32:33.42 evergrowmarketing Yeah. 01:32:35.46 Jon _Originality_ai_ Yeah, but yeah yeah, if anyone has any questions um or comments or or ideas um for for the tool, where we're all ears. Find me on LinkedIn um or john at originality.ai. 01:32:47.53 evergrowmarketing Awesome. And all those links will be in the description of the show. So. Well, John, thanks for joining us on the podcast. Everybody else, we'll see you next time. 01:32:57.73 Cody See you.