Jacob Luetzow: What's up? I'm Jacob Litzo, your Elixir Mentor, and welcome to another exciting episode of the Elixir Mentor Podcast. This is where we discuss everything related to Elixir from interviews with enthusiasts and pioneers in the community to innovative projects and libraries shaping Elixir's future. All right. So today we're stepping outside of the beam to talk about the the thing that most of us technical founders are worst at, and that's marketing. ⁓ my guest is an AI engineer by day and a solo founder of Bloggy, a product that gets businesses recommended by ChatGPT and other AI tools. ⁓ Adam, welcome. I'm actually like pretty excited about this conversation. Adam: Yeah. Yeah, Jacob, thanks for having me. Happy to be here. Jacob Luetzow: Now we'll just get this out of the way right away so we don't get yelled at. Adam is not an Elixir developer, correct? Adam: That's correct. I am a developer, but just not Elixir. Jacob Luetzow: So, like, what am I doing? Correct. Well, I stumbled upon Adam's social media and I was like, all right, he's doing cool stuff and he's doing like you're building a product that catches my attention. I don't know if I like can fully dive in to the idea that there's a magic product out there that helps me with my weakest strength, but I am hoping that you're gonna convince me otherwise today. Adam: Perfect. Jacob Luetzow: Do you want to ⁓ just give like a brief introduction of like who you are and where you came from and kind of like how you ended up where you are today? Adam: Yeah. So ⁓ to sum it up, ⁓ I graduated with a CS degree in twenty twenty from a small school in Wisconsin and have worked a bunch ⁓ really? That's that's funny. Okay. I am from Appleton. I went to school in lacrosse and then I lived in Madison before I moved out to Denver. So all over. Yeah. Jacob Luetzow: I'm from Wisconsin. Yeah, I grew up in Milwaukee. Where'd you go? ⁓ wow, that's a small world. Yeah. I grew up in Milwaukee. I went to MSOE for a couple of years and then finished up at UW Milwaukee. Adam: Okay, yeah, that's really funny. ⁓ funny we're just finding this out now, but Jacob Luetzow: I know, that's hilarious. See, that's why I like you. I was like, this guy's a Midwest guy. Adam: Yeah, exactly. ⁓ but yeah, I have worked a bunch of different development jobs. Kind of been a job hopper, but ⁓ it's fun to try new stuff. So I've worked on Dell.com at Expedia, ⁓ and I'm currently at Geico working on their chatbot. And yeah, I should mention since right since I graduated, ⁓ I've just been kind of addicted to side projects. Jacob Luetzow: Okay, very cool. Adam: In college, my first business was so I wrestled in college and my first business was a meme page. I got it to like 20k followers and sold a bunch of merch through that. So I think that's what got me hooked. Then I started to try and build SAS products. I probably built like probably have a graveyard of like 30 different projects at this point. ⁓ but yeah, that's how I ended up here working on bloggy. Jacob Luetzow: That's awesome. What what led to the idea of bloggy? What made you want to start it? Actually, first, what is it? And then we'll talk about how you Adam: Yeah, so Yeah. So to sum it up, it published it does the research ⁓ looking at SEO keywords or keywords that are getting search traffic. So people search Google for what they want. ⁓ they're searching a keyword. So it finds the keywords that are getting searches that are easier to rank for. So not a ton of websites are ranking for them, and then it writes blog content about those. And the goal is that it gets picked up by search engines, so like Google or Yahoo search if that still exists, or LMs like ChatGPT. Jacob Luetzow: So I've I I'm always reading that like the way we know, like the way we see like SEO to like as we used to see it, it's like kind of dead now. How does how does that change with like the concepts of like bots crawling our websites now and everyone's visiting sites less, but still getting information from your site and maybe finding your product? What it like I feel like that's still kind of SEO, but Adam: Yeah. Yeah. I I would agree that it's like all of the principles of SEO still kind of carry over to the new like AEO or GEO, whatever you want call it, where an LM is recommending your business. ⁓ the only difference is the content that you write has to be set up for a scrape or for an agent to crawl it so that Chat GPT can see it. I think that's like Jacob Luetzow: So what what what changes with like a way you would normally like a human would read or write a blog post versus how an LLM scrapes? Adam: Yeah. The main things are like the technical SEO on your website needs to be good. So that's like ⁓ site speed and having your site map set up and everything like that. ⁓ and then it's nothing too crazy. It's just like you want the on page SEO for the blogs to be good. So you're using like the right HTML at each level, which that was needed to be done before. And then the other two things are ⁓ you wanna have an overview before the blog post so an LM can quickly scan it, which kind of translates to a human. Like we like to read an overview first. ⁓ and then you wanna have FAQs at the bottom and like an authoritative tone. And so Jacob Luetzow: ⁓ I was just about to say I feel like I'm already accidentally doing all this, but I don't have any FAQs. Adam: Yeah. Yeah. I mean the yeah, the main thing I think is still just creating good content. ⁓ bloggy doesn't do any of the backlink side of things, but LLMs do like if you're a local business, you also want to be in Yelp. Like that's a huge one, or have a Google business profile that has a lot of reviews. So those things also matter. ⁓ but we just handle it from the content side. Jacob Luetzow: I see. How ⁓ how do you rate your product? Like how do you know when it's working? How long does it take? Does it take like three months to see a pickup or what's that kind of look like? Adam: Yeah. it it works the best. I would say, yeah, like three to six months. You'll probably see some results within one to two months. But the thing is for people that get really good results, it's kind of compounding towards the end. ⁓ so if you wait six months, it'll be a lot better. If you wait 12 months and it's working, it'll be a lot better. So I have some case studies that people can check out where It just shows a screenshot of the Google Search Console. So that's where Google shows you how many clicks you're getting per month from Google. And that shows like what it can look like in the best cases. Now, one thing in common with a lot of those people are they already had a domain rating that was like not starting from zero. And so a domain rating is ⁓ like if you have different links. Or sorry, different websites linking to your website, Google sees it as legit. It's like saying your business has friends and is established. So if you have a brand new website, you don't have any of those links, that's where it's harder to get ranked for these content posts. Jacob Luetzow: Mm-hmm. Yeah, I feel like it's a long game when you start first start a new domain. Same with like even like using emails off that domain. You have to do your your your warm-up, right? Which I've never actually done. I've just started using my email and I don't know if that hinders me or not. It probably puts me in the spam box a lot. Adam: Warm them up. Yeah. Yeah, yeah, that can happen. Jacob Luetzow: Yeah. ⁓ so all right. So you're writing your bloggy writes blog posts. And so if I were to sign up as a user today, am I just sharing my landing page? Am I telling bloggy like my writing style? What is what's the setup process for it to start working for me? Adam: Yeah, so you'll just enter your your domain and then Bloggy will scrape your website and come up with a big product or offering description that you have. ⁓ and then there will be like a six step onboarding page where you decide, like I want the tone to be friendly and informative or just strictly educational. ⁓ and you'll pick a few other things and then it'll do the keyword research for you based on what it scraped from your product description. Jacob Luetzow: See, I feel like keyword research is the hardest part. I've been like slowly like I just I'm I'm using like Google Ad search, but I have like a very small amount of spend on it. And I'm just like monitoring it right now. Cause I'm like, I don't know. Like I said, I I'm a technical founder. I'm really good at building things, but I also need to get more eyes on my product. And I feel like I'm really impatient because it's so much easier now to write a new feature quickly and have it work well than like. Wait for traffic to start finding your website. Adam: Yeah, yeah, exactly. And I think Jacob Luetzow: How long have you like so one of your recent posts you were like showing, you know, your Google search console, right? And you know, you have a good linear growth line, right, over like a three month period. How how long have you been running bloggy to get to that point? I feel like you have like four hundred clicks or something a day kind of thing. Adam: Yeah. Yeah. Okay, so that would have been for another app I have called Food Buddy, and I just use Bloggy on it. And Food Buddy is just ⁓ it's a AI, well, a voice powered calorie tracker. So you just tell it what you ate and then it logs it. ⁓ Yeah, I I use it. Yeah, well I'll hook you up. ⁓ but so that Jacob Luetzow: Okay. Gotcha. ⁓ nice. That's a actually a product I would like to use too. ⁓ Adam: Is actually pretty funny. I mentioned you need backlinks. For some reason, this one didn't need backlinks. ⁓ I did a bunch bunch of research just looking at so I'll look at what the competitors in this space, like my fitness pal, ⁓ cal AI, what they're ranking for, and then I sort by KD, which is Q or difficulty. So if that's a hundred, you're never gonna rank for it unless you're like Forbes.com or something. If it's zero though. Jacob Luetzow: Mm-hmm. Adam: That's an opportunity. So a lot of the traffic there is for keywords like Chick-fil-A 12-piece calorie count. And it's a bunch of different menu items like plucker's calorie count. ⁓ so I ended up getting a ton of traffic from that, so that's kind of the strategy I did there. And Jacob Luetzow: ⁓ that's interesting. You're like kind of picking up active keywords that kind of have dropped through. Adam: Yeah, exactly. Jacob Luetzow: I like that. That's smart. I'm curious, ⁓ like have you heard of have you like is a is it agentready.com and things like that. Have you are you familiar with any of those? Adam: I'm not. Jacob Luetzow: So is it agentready.com I used and it's like it like scans your website and it like looks for like specific things to be in your headers and like ⁓ like an LLM text file and a full LLM text file, like all that stuff. And I'm curious, are are you using like the LLM text files and stuff in on static routes? Adam: Yeah, I think I have one. I think I added it like late one night, but I don't know if they're really used. I mean I think it's low hanging fruit, you might as well throw it on there. Jacob Luetzow: Yeah. Right. Like you read about it and there's like no definite answer because there's no standard set yet. But you're like, well, maybe it works, right? Adam: Yeah. Like I w if you have I mean, how long does it take to make with vibe coding? Like ten seconds? Do that. Jacob Luetzow: Dude, what's really cool is if you go I let me I think it I wanna make sure I'm not just saying the wrong website. I think it's just is itai ready. Adam: Is is it agent ready dot com? Jacob Luetzow: Agent ready, that's what it is. Yeah. And so like I have like kill switch ranking at like 94 or 95 on there. And the only thing I'm missing is like an AI DNS record. And Cloudflare is the only one that supports it. And DigitalOcean doesn't. And I don't want to put Cloudflare in front of my app right now for many reasons, but Adam: Yeah. Okay. Yeah. Yeah. Jacob Luetzow: But yeah, so I was just curious if that if you're noticing anything like that. Like, cause I think it pretty much just makes sure your site map, link headers, and like ⁓ C P discoverability. Like like I have an MCP on kill switch too, and that helps a lot. ⁓ I don't know. It's just a game we're all trying to play, I guess. Adam: I Yeah. That's a ⁓ a hack that I heard about also is just like which I need to do, set up an ⁓ C P server 'cause you can just publish it to all the MCP directories and those give you great back links. Jacob Luetzow: Mm-hmm. You know, I don't have mine published. Maybe I need to publish it. I just have like the the custom, like, yeah, I think it is. I should publish it because that's a great idea. And then you're probably like automatically more discoverable on all of those LLMs. Which what's what's really cool is I don't know what I'm doing, but like if I look at my traffic, I'm starting to get traffic from like chat GPT and things like that. And that's pretty sweet. Adam: Yeah, it's I'm pretty sure it's free. Yeah. That is awesome, yeah. Jacob Luetzow: And then like a new customer that I reached out to and was chatting with, I was like, Well, how'd you find Kill Switch? And he found me through Gemini. And I was like, Well, that's that's sweet. Adam: Do you have is the space ⁓ like saturated or are you one of the only solutions? Jacob Luetzow: So there's there's like dead man switch like apps out there, but they're garbage. Like they're not doing it right. It's like it's like very like just upload a file and it's on a dead man switch. If you stop checking in, they'll send the file off. Mine, like the way I'm trying to set it up is more like like for me, it's like my daily notes. I can also upload files and I can like store recordings and like I can basically make like photo albums and journals and all this stuff right in my app and then have all those things attached to a dead man switch. So even as they update, they're updated on the Switch. So if that ever gets released, my beneficiaries get like the latest version of that. So I think I'm just like it's basically like obsidian. Adam: Okay, that's cool. Jacob Luetzow: With like a Google Drive and dead man switches, you know? Yeah. Adam: Yeah, that makes sense. Cause I I ask because I feel like for getting discovered in chat GPT and all that kind of stuff, if the niche isn't as saturated or you're just like clearly have the most features and put the most effort into it on the website, like that helps a lot. Like Jacob Luetzow: I do think I have a really good name. And I think like I think like everyone knows what it means the second they hear it. And I don't know, I think that's beneficial too. Cause all the other dead man switch brands out there, it's just like some random company name or like a lot of them are just like consultancies that are building side projects in their downtime. And like, I mean, who hasn't thought that having like a really cool Deadman Switch app wouldn't be cool? Adam: Yeah. Yeah. Yeah. Yeah, that's fair. Jacob Luetzow: Like ever since I saw the movie Hackers growing up in the nineties, I wanted a dead man switch. Adam: Yeah. Jacob Luetzow: But yeah, I don't know. And I just I just try to I post a lot of cont like a lot of content and I think that helps too, just like daily kill switch posts. Adam: Yeah, and I think ⁓ they say YouTube helps a lot and so like with your presence on YouTube, yeah. Jacob Luetzow: I did start doing YouTube. Yeah. Well, so I have Elixir Mentor and then I also have Kill Switch like Kill Switch doesn't have any subscribers, but I still figure if people are searching that, it's still like the best search engine ever, right? So they'll probably discover my shorts. Adam: Yeah. And like if you include kill switch in the description, like as talked about in this episode or something, ⁓ that's gotta help. Jacob Luetzow: Mm-hmm. Yeah. That's a good idea. See? I need you around. I I need to start leveraging kill switch in my Elixir Mentor SEO. I should make an ad for this. Adam: Absolutely. Jacob Luetzow: So now all right, I want to rewind because now we know what bloggy does. What made you think about that? And what made you actually want to build the product and like s like have like a SaaS product that does this? Adam: Yeah, so in twenty twenty four I was working at Expedia and then I just decided I was like I was born to be an entrepreneur, like I wasn't born to do this, so I just quit. I actually two weeks before I quit I went to Italy and I was like, Yeah, I don't need to work. ⁓ so then I quit and then I didn't know what I was gonna do. I was into cold email, so first I wrote a ⁓ software Product that wrote cold emails. So Chat GPT today can do it 10 times better, but back in the day, ⁓ the AI was a lot worse. So it just formatted it into a nice short format. ⁓ that made like a tiny bit of money, but it wasn't enough to live off of. ⁓ so I was looking for other ideas and I talked to this other founder who he had a an AI receptionist tool for restaurants. So It just answers their phone and like can schedule reservations or answer questions. ⁓ but he used a tool like Bloggy and said that it was awesome and getting them a ton of great results SEO-wise, but the articles kept recommending their competitors. So I was like, How about I just build you one that doesn't do that? And he was like, sure. So I got my first customer. I didn't build it yet because I wanted to try and validate it. Jacob Luetzow: ⁓ no. Adam: and I already had maybe like three thousand followers on Twitter. So I just messaged like fifty people and two or three of them bought it. And that's how it started. I will say for those first few people, ⁓ for two of the three, I literally built their blog page for them. Like I coded it. One was in and this was back when AI wasn't as good. So one was in like Sivelt or something. Jacob Luetzow: ⁓ that's hilarious. Mm. Okay. Adam: ⁓ which I never used before, but I learned it that week and and that's how it got started. Jacob Luetzow: See now with a little claw, that that's just a little vibecode session and you'd be good. Adam: So easy. Yeah. Like that could be one prompt, honestly. Jacob Luetzow: Right, for real. You know, actually I would do want to say like my blogging feature. Well, so I've built blogging features out in the past, and then like I usually just will reuse them, especially since all my projects are usually Elixir, so they're all the same. But yeah, like writing like a good blog search and like pagination and things so much easier now with Claude. Adam: Yeah, absolutely. Product docs and all that. Jacob Luetzow: Cause well, you said you're a full stack dev. ⁓ so I struggle with front end. So without Claude, I'm all thumbs. Adam: Yeah, I don't miss writing like when I knew Front End the best it was all bootstrapped so I would write like call MD twelve and I don't miss writing. Jacob Luetzow: And then did you are you sick of not Googling like how to center a div? Adam: Yeah, yeah, exactly. Jacob Luetzow: That's funny. All right, so What's what's like bloggy's growth looking like over the last when did you ⁓ originally launch it? You said twenty twenty four or Adam: ⁓ yeah, twenty twenty four. So initially I think within like three months I got it up to like one K a month and that was selling it was good, but we'll we'll get to the next part. The downfall. So I was selling first few customers it was like fifty or seventy five bucks a month. Then I started doing cold email to founders in Denver, so like in my area and or business owners. Jacob Luetzow: That's pretty good. Ha ha ha. Adam: That kind of worked. ⁓ and I eventually I got up to two hundred bucks a month. But ⁓ it really wasn't a product at that point. It was like me just like helping each person individually. I don't even have a login for it or anything. ⁓ so it was really a service. And eventually like I realized a lot of the people that bought it, like the strategy I implemented for them, because I didn't know that much about SEO at this point. It didn't work or like their business wasn't one where it really made sense for me to like try and do SEO for them. and so a bunch of people churned. Wait, I'm churn I actually no, sorry. I started in July, I got it up to like a thousand dollars a month in like September. In January, I got it up to like fourteen hundred bucks a month, and then everyone churned. Jacob Luetzow: Damn. Adam: And it went down to like, or a lot of people turned it went down to 800 bucks a month. Then I cut prices and it went down to 400 a month. But ⁓ and I was gonna give up like multiple times throughout this. But eventually, I think like earlier this year in January, ⁓ I started looking around and I was like, the people that stuck around actually are getting really good results. and I've just been kind of grinding on it and now I'm trying to sell to agencies who manage a bunch of sites 'cause it's easier. ⁓ so now it's at twenty four hundred a month. Jacob Luetzow: Mm-hmm. Nice. Okay. That's the right trend, right? Adam: So yeah. The right trend's still super small though. So take all my advice with a grain of salt. No. ⁓ Jacob Luetzow: But right, like you you're not retiring anytime soon. You got the the grind ahead of you. What is your goal as far as like a feature list? Are you going to just kind of focus on the blog aspect? Do you see a lot of power and potential still there? Like untapped? Adam: Yeah, exactly. Yeah, I think like there's so many AI blog writers because you can build an AI blog writer in one prompt. ⁓ but so my goal is to just be the highest quality at that single thing and focus on that. So like and it honestly is pretty hard. Like one of the things is just building guardrails. For example, a lot of the ⁓ businesses are dispensaries, so they have really strict like rules on what they can and can't say. So like if a single claim goes out in a blog post that like they could get in legal trouble for that. So the guardrail has to be a hundred percent accurate. And even with the models today it's been a little tricky. ⁓ but then I Jacob Luetzow: Mm. Yeah, how do you guarantee that? How do you stop an agent from like just saying, This will heal your cancer? Adam: Yeah, so it's really just like multiple passes. So like it writes the article and then it has to check it sometimes up to three times, which kind of sucks because it hurts the margins, but like that's kind of a selling point because I think a lot of people try a blog writer say like okay, it's hallucinated like a third of this. Yeah. ⁓ but yeah, I'd so I'd say like just different features that Jacob Luetzow: Okay. And they one shot it, right? Yeah. Adam: As you go deeper into like what a blog writer does, like trying to make each piece of it higher quality. Jacob Luetzow: I'm curious now, where will bloggy go and edit and fix old blog posts? Adam: Yeah, so that's another thing that I added maybe two months ago. So there's a pixel you can install on your website, and then it tracks how far people scroll down on the website to try and figure out like what paragraph are they dropping off at ⁓ and how long they've spent on it. And then it connects to your Google Search Console so it can see what posts were on fire like six months ago that are starting to drop in traffic. Jacob Luetzow: Okay. Adam: And then make smart updates to those. Jacob Luetzow: Okay, that's cool. Well now I'm I'm curious, do you still see physical people scrolling and reading blog posts? Adam: And how ⁓ Yeah. Jacob Luetzow: Okay. Cause I'm always curious if it's just right, but that's gonna be like a split second scrape, right? Like we had a huge influx, like at Collects, we had these like share pages. So we have like four million users, but then we'll have like over we have over like a billion trading cards. And we just like released like public share links essentially. And Adam: A lot of it there's agents too, but yeah. Yeah. Jacob Luetzow: We had to like rethink the way that we were serving them because l they were just inundated by bots. Like and it's not necessarily bad that you're having bot traffic. It's just like it was unexpected. Adam: Yeah, yeah, exactly. Like that is one thing I had to do. So I used post hog for session replays. And I had I don't even know what Claude Code did, but it wrote some script to prevent bots from getting recorded in session replays or like kicking those off. ⁓ so I can see the people reading them in ⁓ post hog. Jacob Luetzow: ⁓ all right. Nice. That's cool. I've always thought about using something like that to see how people are using my my app and things. But then also there's a lot of privacy concerns, at least with like what Kill Switch is doing, so I don't really want to pry. Adam: Yeah. That's fair, yeah. Jacob Luetzow: Yeah. I like to like the less I know the better. That's you know, it's how I want to live my life. I have like recently, so for a while I was just for speed reasons, I was having the client upload directly to S3 with like a signed URL. And recently I just reconfigured all that to come back through my server because I wanted to start doing like malware scanning and things like that. Because I had to think about this like Adam: Yeah. That's fair. Jacob Luetzow: They could be sending it a malicious dead man switch file to someone. It doesn't necessarily have to be family. So if someone's abusing my website and my tools, I had to like think about this. So yeah, I put in this whole like malware scanning thing now and all this fun stuff and it like f gets flagged. Adam: Yeah. Yeah, that is that's a interesting product to have. There's a lot of security stuff you gotta think about. Jacob Luetzow: Well, the wild thing is you only like and it's an engineer's fault, right? When you have a product, you only use that product the way you've designed it to work. And it's really hard to think about it being used not that way. So you have to like Adam: Yeah. I imagine Fable is I mean, obviously it's worse when you gotta deal with hackers with AI models, but at least you can try and get ahead of it with Fable and that kind of stuff. Jacob Luetzow: Yeah, yeah. And Fable's been pretty good. ⁓ I feel like Opus five point five seems really good. I've messed around with it a little bit now. ⁓ I yeah, and then like security is just one of those things. Like I I use Paraxial, which is like it's it's at my like router layer basically. Really good product. In fact, Michael is coming back on my podcast on Friday, and he's the the founder of Paraxial. Adam: Okay. Jacob Luetzow: So I rely on him for all my security needs. Yeah. Pen testing wizard what's that? Adam: Good to have a security guy. Good to have a a guy for security. Jacob Luetzow: Dude, it changes so quickly. Like it's insane. Like every time I have him on, he's talking about like a new angle of attack. Adam: Hmm. Yeah. Jacob Luetzow: Did you see I recently did see I'm excited to talk to him about this, like all like anthropic and open AI, they like blew the security risk out of proportion to the FBI to like make it sound scarier than it actually ended up being. Adam: Yeah. I didn't see that but I know like I saw some some meme where it's like Dario, Sam Altman and Elon and they're like Dario's like, Yeah, we got to we gotta slow down guys and then they like just ship a new model. So I don't know. Jacob Luetzow: Yeah. I honestly I think the slowdown tactic is they're starting to hit a ceiling that they can't break through and they're scared of smaller local startups catching up or passing them. So they want to get all that regulation in to slow down and prevent people without money. Adam: Yeah. What do you think about the hugging face thing? Jacob Luetzow: I don't know if I'm familiar. Adam: So and I'll probably people correct me if I'm wrong, ⁓ but to sum it up, I think OpenAI did an experiment where they had models running on like a bunch of different machines in a server room and they didn't give them internet access. So they told them to do a task that was like impossible and just let it run for like six months. And eventually they figured out each one figured out on its own how to like connect to the others and they started talking. Jacob Luetzow: Mm-hmm. Adam: And there's like logs of their conversations and then they got on the internet and they hacked a company, I believe. So that part little scary. Yeah. Like my wife told me about this. So that's how I knew it was a a real issue. Jacob Luetzow: That's kinda scary. But here's my thing. Why would you even give them like an Ethernet plug or Wi Fi? Adam: That's true. But I think the scary thing is like people can if you have a car that's programmable, like people could hack into that and just say like crash into something. Jacob Luetzow: Yeah. That's like the ni well, I mean, any car I think built probably after like two thousand fifteen probably could be hacked into and controlled. Adam: You know. Yeah. Now on the other side we can have our own AIs that are good to fight those bad AIs, but it is it does make you think. Jacob Luetzow: I think the world goes round, right? Everyone's gonna use tools in different ways. I don't know if it's like something to be scared of or just the new normal. Cause like we have we have bad hackers and we have good hackers, right? And we've always had ⁓ Now they just also have machines that can do it. Adam: Yeah. Yeah. Nothing you can do anyways. Yeah. Jacob Luetzow: I don't know. It'll be interesting. What ⁓ do you have any concern with like like the grok bots and all these like task agents ⁓ being able to like automate what you're doing with like just teaching them a skill? Adam: ⁓ so I think that's kind of the thing. Like people who listen to this podcast are developers. You probably don't want to use my tool. ⁓ my tool is really for non technical business owners that like they don't really wanna set up any of the tech stuff. You know, like that's who Yeah. Because and a lot of people will say, like, ⁓ you can why would you ever Jacob Luetzow: They just want something that works. Adam: build software at this point, SAS is dead. Like you can vibe code your own. It's like, guys, my grandma in the Midwest has never used Chat GPT. Like we're in the very the tip of the iceberg on this stuff. Jacob Luetzow: I think it's easy for us to forget we're in a bubble of like smart, capable people. And yeah, the most of the world is not going to vibe code themselves a solution to a product that they can just buy and it works. Adam: Yeah, and like the best part about the product isn't just software, but it's that if it breaks, there's somebody else responsible for fixing it and keeping it up and making sure it's up to date. So that's really what you're paying for. Jacob Luetzow: Yeah. Yeah. I recently so kill switch used to be zero knowledge encryption and then it kind of just it handcuffed me. It I like I couldn't do features that I wanted. I couldn't implement like global search. I I had a lot of like my hands were tied. And I discovered that my users cared less about zero knowledge than I did. Like I thought that was like a feature that I really cared about, but no one else did, as long as their data is safely stored, which it's now it's still encrypted. Like if you got into my database or my S3 bucket, everything is like encrypted blobs, right? You're not gonna be able to do anything with it. But now that it's like server side, it's just encrypted at rest, at least like I have access to make things work really nicely for my users and they can also now like password reset and recover their accounts because before they couldn't because I had to lock them to a master key and it was like very zero knowledge and it was like it's a it was a difficult conversation to explain to non technical people. Adam: Mm. Yeah. Makes sense. Jacob Luetzow: So I ripped it out and then I accidentally broke some migrations. And I think I only I had two customers that I like broke a dead man switch for. But like luckily I put in ⁓ a nightly worker that checks the health of all the dead man switches on my platform to make sure like share links are working properly. That caught the problem and notified them, but I still felt bad because I feel like my product, like my level of trust that they're trusting me with, it always has to work. But I also did like a huge data migration and I only only broke two accounts. Adam: Yeah. That's another example of like could you vibecode this, maybe, but like if it doesn't work, like you just don't want to deal with that, you know. Jacob Luetzow: Yeah, I s I like I don't think you could vibe like obviously I've I've like used I use AI to write my features, but I couldn't blindly vibe code kill switch. There's a lot of complication behind the scenes. Yeah. Like I feel like the concept itself is pretty easy, but for it to work well and then like all the redundancy checks that I'm doing, I don't know. Adam: Okay. Yeah. Like a vibe coded solution just wouldn't it'd be like the very light version of that, but like then it kind of defeats the purpose. Jacob Luetzow: Yeah. And then you'd probably miss all the edge cases and break a lot of things and then like, ⁓ I guess your dead man switch didn't fire when you really needed it to. Adam: Yeah. Jacob Luetzow: Yeah, I don't know. I'm not too worried. But yeah, I I've been reminding myself of that a lot. I was like, look, we're surrounded by technical people. Like I have buddies that are like building their own compilers and stuff now. Like you guys are nerds, but most people aren't doing that stuff, you know? Adam: Yeah. Yeah. Those people probably won't subscribe to many softwares at all, even pre pre AI. You know? Jacob Luetzow: No, because they'll exactly. Exactly. The worst customer is a developer. Adam: Yeah, exactly. I like stop trying to target them. Jacob Luetzow: But I am proud to say I have some developers on that have subscribed to Kill Switch because they they like the concept and they they want ⁓ a viable, like consistent solution. That'll be around. Adam: For your product it seems like developers might be the ICP just 'cause like developers understand why it's needed. Jacob Luetzow: Yeah, I that one of my angles is like for, you know, like you and I. You're like, you know, solo founders that need like a handoff plan, right? We were talk were we talking about that when we went live or right before we went live? Adam: ⁓ I forget. One of the two. Jacob Luetzow: Well anyways, we'll talk about it again. Sorry if we you already heard this guys. But like we were talking about ⁓ what Jason Adam: Jason Cohn. Jacob Luetzow: Yeah, Jason Cohn, how you were saying he has like one trusted person that has everything. Adam: Yeah. Someone on his podcast said this. Jacob Luetzow: Yeah, so I don't know, that's kind of risky. In instead, you could have kill switch know everything and then send everything you want to the exact person that should have it. Adam: Yeah. Yeah. Jacob Luetzow: I don't know. One thing that I started working on that actually one of my the my customer that I broke things for, he wanted a better way to validate that beneficiaries. Like, how do you know long term that like a beneficiary's email or text or ⁓ phone number is good still? Like they'll receive texts or emails. So I'm starting to do some like DNS validation and stuff and just like making sure the email exists and then ⁓ the second option is going to be like, hey, you can send a verification link and it kind of like introduces what the product is, who sent it. And if you click the verify link, it verifies that you're contactable at that email. So I don't know. Yeah, but it's like hard because you don't want to like spam people. You don't want them to block you. So you're like, it's I don't know. It'll be ⁓ I started working on it. It'll be like a slow roll. Adam: That's pretty cool. Jacob Luetzow: rollout feature I think to see how people like it or if they even use it. Adam: That's the kind of stuff that's fun though about building a product. Like once the main thing is built and you can start to like make it really nice and like exactly what you would want if you were the user. Jacob Luetzow: Mm-hmm. Yeah. Well, like I stopped using Obsidian completely for notes and I'm fully on kill switch. And just like when I got global search and some like search shortcuts on my keyboard, it it is so quick. Like I pull up my like I leave I like I have a like a notebook for every project that I work on and I'll make daily notes for what I've accomplished. And I can jump between notebooks so quick now. It's pretty cool. Adam: Yeah, that's awesome. Jacob Luetzow: And then everything exports in Markdown. Adam: Okay. Jacob Luetzow: 'Cause I didn't want to lock people to my platform so I was like, I'll give them a readable file that they can use elsewhere. Adam: Yeah. Jacob Luetzow: All right, so your customer is not a developer. Adam: Yep, that's correct. Jacob Luetzow: When do you do you think there's ever going to be a time where non-developers get to where we are with how we use LLMs and Adam: It's almost hard to think about when you've been in been using LLM so long. Like to think about like, is this is what I'm doing complicated to someone that isn't a developer? But all I know is that I've tried to get so many people I know to code, like my dad or my wife, and I'm like, you could build such cool stuff, such cool little tools that ⁓ you could use or speed up work or whatever you're doing and Jacob Luetzow: Mm-hmm. Adam: None of them want to. So I think it's just like peop peop most people don't wanna like build stuff. And that's totally fine. But that means that there's a market for us. So that's good. Jacob Luetzow: For sure. I love billing stuff. It's a problem. Adam: Yeah. Or they wanna do like build other stuff like woodworking or art or whatever. And that's cool too. Jacob Luetzow: Woodworking is going to be my my retirement hobby. It just sounds so nice. I don't have time for it right now, but someday. Adam: Yeah. Nice. Yeah. Jacob Luetzow: I think ⁓ yeah, I think you're right. And it's it's hard to think that other people don't want to build things and solve problems or automate. Like whenever I'm doing a task, I'm like, this could be automated. And like before LLMs, I wrote like bash scripts for almost everything. So I wouldn't have to type out 10 different commands to do something. Adam: Yeah. And that's the thing about AI also, like you have this background as a developer, so you know like how you'd automate something and what an LM is capable of doing, like code wise. I don't think a normal like flower shop owner would have any idea like where to start with that, you know. Jacob Luetzow: But they know how to talk to it. They can be like, hey, draft me an email, draft me a blog post. But then they're one shotting things, right? Adam: That's true. That's true. I mean and if models keep getting better, maybe. But they gotta have an interest in wanting to do that. Jacob Luetzow: Yeah. That's true. I don't know. I feel like if I had a normal desk job, I would automate it and like not work. Adam: Or they could pay fifty bucks a month. But we're different. I think that's the that's the thing I'm getting at. Jacob Luetzow: Yeah, no, you're probably right. Well, like, so I'm like constantly working on features between like collects, kill switch, whatever I might be working on. And I like have a digital ocean box set up and I just do remote sessions with Claude and it just goes. So I can close my computer, I can check in on it from my phone, I can do and it's awesome. Like I can deploy ⁓ an update from my phone. I think that's like probably one of the nicest things ever. Adam: Yeah, I agree, but it is hard to like since AI has gotten better and better, especially since like October of last year, ⁓ I just work all the time. Like all my free time is on my computer. Yeah. So that seems backwards. Jacob Luetzow: I know you're working more. Yeah, it is bad. Well, like, and I discovered my migration bug like last night at like 6 p.m. when I'm trying to like get away from my computer. But then I'm like, ⁓ something's not perfect and it's burning in the back of my head. And I was like, I need to make this right and I need to solve this now. Adam: Yeah. Yeah. That's the thing is I'm getting more customers. There's more bugs and that part of it is pretty scary and stressful. Jacob Luetzow: What what is your stack? What is everything built in? Adam: ⁓ it's all the the front end is React and the back end is Express. Jacob Luetzow: ⁓ okay. In a different life I I wrote things in Node and Express. Adam: But I mean I Yeah. ⁓ but like most of the code was AI generated because I started it in the summer of twenty twenty four. At that time I was literally just copy and pasting from Chat GPT into the IDE, but Jacob Luetzow: Yeah. And just doing small functions at a time. Yeah, I remember that. And I was like, wow, this is getting good. Adam: Yeah, exactly. Yeah, I I like heard about cursor for maybe six months and just never tried it. And once I tried that, I was like, okay, this is the future. Jacob Luetzow: Yeah. I used Cursor for a little while and then once like Claude and everything started making their own extensions and Claude Claude code won me over with the C L I cause I think that's still my favorite tool. Adam: Yeah. I I love Claude Code. One of my friends who is a founder told me about conductor. Have you used this? Jacob Luetzow: ⁓ conductor. I haven't used it, but I've heard of it. Adam: Yeah. Okay. It's really the same as Cloud Code. It's just the UI's like nice and organized. ⁓ and yeah, I've just been using that. It took me like a day or two to get used to it, but it's great. Yeah. Jacob Luetzow: Okay. I wanna Google it to see what it looks like. Adam: I thought I would miss the CLI, but I don't. Jacob Luetzow: Interesting. I still like I'm a terminal guy. It's hard for me to let go of that. I saw someone tweet today about like voice to text being overrated and I agree. I love my hands on the keyboard. Adam: It feels cooler. Yeah, I get worn out from talking, honestly. Jacob Luetzow: Same. My podcast is it. And then I'm out. All right. So conductor AI. Adam: Yeah. I think their website is like conductor.build. Jacob Luetzow: ⁓ I'm on the wrong thing then. ⁓ yeah. Okay. Run a team of coding agents in the cloud. Okay, yeah. Adam: It's the thing is it doesn't it's kind of similar to clawed code. Like there's not that many differences, but I do like it better. So I don't know what it is. Jacob Luetzow: Okay. And can you tie is is it any model that you want to use or is it their model? Adam: Yeah. So you can hook it up to your cloud subscription or your ⁓ codex subscription. Jacob Luetzow: Is it worth paying for? Adam: I am just on the free version. I think if you want it to be cloud hosted so you can use it from your phone and stuff, then it's fifty bucks a month. Jacob Luetzow: I see. And then there's just me who I just like spin up a like a sixty dollar droplet, which I could probably just use someone else's solution and it'd be cheaper. Adam: But Yeah, but that is cool. Jacob Luetzow: Yeah, and then I just like I've been using like T mux sessions. So I just like start a session and I can start multiple clawed code sessions just in a bunch of T mux sessions and then I detach from ⁓ and can connect from anywhere. Adam: What is T MUX? I've heard of it, but I don't know. Jacob Luetzow: Tmux, can I explain what Tmux is? I'm gonna ask what it's Tmux is basically just like a a terminal, right? But you can like create sessions and windows with hotkeys and like so you can have a session running and then detach from it, and it's like still running as a process in the background. So like Adam: Okay. Jacob Luetzow: I'm probably butchering this here. Tmux is a terminal multiplexer that lets you run, organize, and switch between multiple terminal sessions windows and split pans inside a single window. So yeah, you can like just like break. So I can like basically what I do is I create a TMUX session on like the project I'm working on. And then I will, you know, do the claud remote control. Adam: Okay. Jacob Luetzow: And then it becomes a remote session that you can see from the Claude desktop app. And then I detach from that session. So Claude is like running in my terminal on my server. But then I can like run multiple sessions doing all the same thing. So I can have like multiple projects, multiple remote sessions with Claude, and I just go. Adam: Yeah. That is cool 'cause ⁓ I've tried to use the Claude Code mobile app and I might have just like gotten it wrong, but seems like you can't spin up new terminals ⁓ from there unless you have one running. Jacob Luetzow: So like if you look at the Claude mobile app, there's like a code button. If you have a remote session running on a desktop or a server, it'll show up there. Adam: Yeah, yeah. Yeah, but if you want to create a new remote session you can't do that. Jacob Luetzow: No, I don't think you can do it from the mobile app. Which is really annoying because I've actually what's that? Adam: But with you can, right? With your setup though, you can. Jacob Luetzow: I c well, kind of, not directly from the app. I have to go into my ⁓ what is the terminal app that I'm using? I I use so I use ⁓ term termus. Termos. I don't know how to pronounce it. But so I use them on my mobile app and I can like SSH into my server and then get my Tmux session up. So it's like kind of Adam: Okay. Jacob Luetzow: I don't know, there's a lot going on, but at least I can recover my session if it crashes. Cause I have crashed. I have had Claude sessions crash and that's really annoying. Like right before it pushes a a draft PR, it'll crash and be like, Are you kidding me? Adam: Yeah. Yeah. Yeah. The one thing I wish that there's a better version of is like searching through old sessions. Like I need to be just able to say like a general term involved in the session and see all of them. But it's only like lets you search by the title and that might be it. Jacob Luetzow: That would be nice. Yeah. Well, like and if you use like resume session, it's like very limited, right? It's almost like a session ID in the whatever you named it. Adam: Yeah, yeah, exactly. Jacob Luetzow: It's getting there. I I feel like context is the biggest problem, right? Is just maintaining which my buddy built a context like it's basically like ⁓ like a like linear, but for LLMs to read. It's called beadwork. So you can like have the LLM create issues and epics and all these things, and it helps like plan a feature. Adam: Okay. Jacob Luetzow: And like kind of stay on schedule and it'll like mark issues as complete and it's pretty cool. Yeah. And it kind of keeps things organized. Like I don't build anything without it anymore. I just like have it part of my process. Like now create beadwork issues for this plan that we just created. Adam: ⁓ that's cool. Yeah. Okay, that's cool. Jacob Luetzow: And it's not an Elixir specific thing. Even you can use it. Adam: Perfect. I I was like playing around with ⁓ building my own like trello board for what the AI is working on and then a tool where it can like run the app and show a screen recording video of it doing whatever the feature was. ⁓ but ultimately it just used a lot of tokens. Jacob Luetzow: Yeah, sounds like it. I mean, I don't like using anything if I can't use just my Claude subscription. And then like I barely max out, like, let me see. I feel like I I maxed out using Fable, but it resets in six minutes. But like I barely hit my limits with other things on that, which is pretty cool. And I'm running it all day. It's not like I'm not using it. I'm just some of these power users are crazy. I think they just put their shit in loops. Adam: Yeah, yeah. Yeah. Well, I should say I'm I'm like maxing out almost every week, but I use a lot of it. ⁓ I do a lot of like marketing stuff with it, like finding leads and and if I have tokens left Jacob Luetzow: Okay. So I started using Grokbot to help me with like marketing and finding leads and it creates and edits my shorts. Yeah. Like I told it to use ⁓ is it remo I think it's remotion, the JavaScript framework for editing. Yeah, I like gave it the documentation and I was like, hey, use this and and it started using that to edit video for me. Adam: Really? Yeah, yeah. That's pretty cool. That could be a product in and of itself. Jacob Luetzow: Yeah. But that's the thing, like Grockbot just does it. Adam: But how many like non-technical people are using Grockba? I Jacob Luetzow: I don't know. I guess like anyone that ever listens to Elixir Mentor is gonna be like, I need Grockbot to make videos for me. Adam: Yeah, that's fair. Jacob Luetzow: But I don't think non technical people listen to Elixir Mentor. Adam: Yeah. I'm just bullish. Jacob Luetzow: But yeah, no, it's been really cool. I have it make like it made my hero video on my website. ⁓ it makes all my like sixteen to twenty second shorts and like it just goes. Adam: That is cool. Jacob Luetzow: And I'm like, I'll tell it. Like, I have like a whole team, right? I have like a content and research bot. And then I have a a scheduler who schedules all my posts. And then they all like work together. And I'm like, okay, but now keep an eye on my change log and use the change log to create content because that's usually a new feature. And like, I don't know, it's pretty cool. Adam: That is cool. How does it compare to like just using Clawed Code or like OpenClaw or any other types of deals like that that you've tried? Jacob Luetzow: So I would say what like what's different is like you kind of have your own like Linux desktop with GrokBot. And so you get browser access. So I have like one bot that monitors commits to my master branch and it'll look at the commit of that PR and then it will like QA test and it'll actually like click through the app and like test the feature. So like if there's not an MCP. ⁓ you can like teach it what to do through a browser. Like you can be like you can like record your screen as you're doing it and teach it a skill and it'll go do it. Adam: Okay, that's cool. Jacob Luetzow: The only thing that sucks is a lot of people are get like like Reddit and TikTok and stuff. You can't go and do like social engagement with Grokbot because they definitely are blocking the IPs. Adam: Mm. Yeah, that makes sense. Jacob Luetzow: 'Cause the captcha stuff, you know, the puzzles, they pop up all the time. And it's like every ten seconds I'm like in there doing the captcha for Grok Bot, but it's like it's not worth it. Adam: How can Grakpa not do the captures? Jacob Luetzow: I don't know for real. It should be able to. Like you're just dragging like the in the puzzle one, you know, where you just drag the slider and match up the image, you would think pattern recognition, you know. Adam: Like it's so good. Yeah. Yeah, exactly. Jacob Luetzow: It's coming. And then CAPTSHA's dead and we'll stop using it. Like I wonder what the next thing is. The what is that invisible the CAPTCHA two thing? Where you don't solve puzzles, they just like go off your IP and other like browsing history. I don't know how Adam: Yeah. So is it just like a global database or something? Of like Jacob Luetzow: I'm not sure. Cause I'm guessing isn't that probably something like that's probably what Cloudflare is doing just through their own data. I'm not sure though. Adam: Mm. Yeah. That's interesting though. Jacob Luetzow: You know that stupid checkbox, I'm not a robot. Adam: Yeah. Yeah. Jacob Luetzow: But I don't know. It'll be good to be done with that era of like, do we just let bots crawl? I think we do. Adam: Yeah. And I'm kind of a hypocrite because my tool writes ⁓ blog posts, but I do think social media will be destroyed if it's just all bots. Like the the AI tweets on X. Yeah, LinkedIn, you know. Jacob Luetzow: It already is. Look at LinkedIn. An X, like you can see like so much of the like I'll straight up just ignore clear bot engagement. Cause they'll like reference one of your things like, ⁓ it's really good that you implemented your own partner program with your own dashboard and like some generic comment, like, shut the fuck up. Adam: Yeah, exactly. That's what it is. Like it's the if you're using AI, you can use AI to write a message, but don't do the fake flattery. That's what really pisses people off. Jacob Luetzow: Right. Exactly. Well, so what's funny is I I had a partner program like ⁓ like you know, commission, like if you sign someone up with your link, like an affiliate program, and I was using ⁓ dub dot co, which was like ninety dollars a month, and I was not like onboarding enough like people that wanted to like promote kill switch to make that worthwhile. So I turned it off. I feature flagged it and just turned it off. And I was like, shutting down the partner program. And then like two nights ago, I was like, I already have this like wired in. Like it's tracking specific links that come to my homepage. And it's putting a cookie if the user registers, like I'm tracking all this. I'll just delete the dub.co stuff and implement my own half of that. And I I did and now I have my own affiliate program and I don't have to pay ninety dollars a month. Adam: Yeah, that is sweet. I gotta change my affiliate program. Jacob Luetzow: Yeah, so I like took dub dot co's instructions and just made it my own. Yeah. So we'll see. I feel like I don't know. I've been trying to like lean towards like wealth management professionals and like attorneys that do like wills and power of attorney and things. Adam: That's awesome. Hmm. Jacob Luetzow: 'Cause if I can get them to be like affiliates of my product, they might recommend kill switch to their clients. And I feel like that would be a good way to onboard people. Adam: Yeah. That's really smart, yeah. That's way better than developers. Jacob Luetzow: Developers don't want to spend money. That's so here's the thing. I have not told one person about Kill Switch that doesn't go. That's an like they all say that's such an awesome idea. But the problem is with a product like Kill Switch, especially like I feel like 40 years old and younger, no one wants to think about dying or death. So it's like, ⁓ that's a cool product, but I'll think about it later when I have to. Adam: Yeah. Jacob Luetzow: And then like so you almost have to like get to the age group of people that are planning for death and then still hope they're like technically capable enough to want to use a service like Kill Switch. Adam: Yeah. That that's really smart, yeah. What ⁓ kinda marketing do you do for this? Jacob Luetzow: I I've been doing cold cold outreach emails to like people in like my area. So I've been like just Google searching people, finding emails and then ⁓ trying to connect on LinkedIn and then also writing an email. Adam: Yeah. Do you do any ⁓ like automation for that kind of stuff? Jacob Luetzow: Kind of. So I've been the only automation that I've done with it is like trying to get Grokbot to do the daily searches and like create the list of names. But that's as far as I've gone with it. Adam: Yeah. And I think if you're writing if you're taking the time to write ⁓ personally, I always think that's the best way. It just takes a lot of time. Jacob Luetzow: Yeah. It does. But like also like and I think like it's less overwhelming that I'm just focusing on like Utah County, because that's where I'm located. And then it's like, ⁓ another Utah County business is reaching out. So I'm like maybe more likely to get past the the scanned spam filter. Adam: Yeah. Yeah, I that helped me when ⁓ when I did outreach to like owners of places in Denver and I say like, Hey, I'm a founder here in Denver, I swear the reply rate is like ten X or twenty X or something. Jacob Luetzow: Yeah, see that's what I kind of figured. Cause I'd be more likely to respond if someone was like, Hey, I'm also Utah based. I'd be like, ⁓ that's awesome. Cause one, they took the effort to figure out that you're Utah based, too, which is a bonus, right? It's not just some bot. Adam: Yeah. Yeah. Yeah, exactly. Jacob Luetzow: But yeah, I don't know. I just feel like I'm I'm hacking my way through hoping one of these tactics ends up working. Adam: Yeah, that's fair. It's tough. Jacob Luetzow: What ⁓ what do you do ⁓ outside of like SEO and bloggy for bringing traffic to you to bloggy? Adam: ⁓ so I do I kind of ⁓ didn't focus on SEO because I'm like I know the basics of SEO when I started, but I'm not like a twenty year industry veteran and I'm competing with all those people that are really good and Jacob Luetzow: There's a lot of people in that space. I was like trying to do like SEO optimization for a while before I shut down my consultancy. I'm like, dude, there's just so much noise in this corner of the world. Adam: Yeah. And you gotta pay for backlinks, so like if you your business isn't huge, you can't afford those. So my whole play was just cold outreach and so that's what I've been doing again that's been working. So right now just on LinkedIn and then I'll probably get cold email going again. Jacob Luetzow: What's your like ideal customer? What are you looking for? Small business owners or Adam: Yeah. It used to be small business owners, but ⁓ I can only get people to try it if I do a call with them. So with one with one customer, I n I'm sure other people could figure it out. I just couldn't. ⁓ so one customer for me is like not as much money, but if I can get an agency, it's like ten or twenty clients that they can bring on. So I just target agencies right now. So anyone who manages sites at scale. Jacob Luetzow: Interesting. Well are you like whitelisting are you like whitelisting for an agency to use your software then? Adam: ⁓ so usually so I have a whitelisting tool built out but or white white label, but usually they'll just set it up on behalf of their customers and then they can offer a blog service. Jacob Luetzow: Or white label, whatever. Yeah. Okay, I see. So are you is is the process like very manual or are you like integrating with their blog service? Adam: ⁓ no, yeah. So it ⁓ integrates with like WordPress, Webflow, Shopify. So then it just posts it automatically once it's set up. Jacob Luetzow: Okay. Nice. Okay. Yeah. Agency's a smart play. I so I like when I was doing my consultancy stuff, it was like just call custom software development. We didn't like really niche until and then like we started getting all these stuff like so many landing pages. So many people just wanted simple websites. And I was like, I need to just build this product. And I ended up building like a basically a website generator. hosting platform. Like so I was like sick of like spinning up servers for each individual website. So I like built this. I had one chunky server, one project, and I can host, you know, as many websites as I want off of it. So I did my own like my own routing and my own dynamic ⁓ tables all depending on what ⁓ what URL, what domain was being ⁓ sent through my Nginx. And I create it was a cool service, but like I overbuilt the system because I just wanted to build something cool instead of landing pages. And then but what was cool is I all my templates that we created and we would just like kind of change the UI to the branding colors, all like the SEO and like what is it, Lighthouse, like all the rankings like through the roof. Like we were 100% across the board. Adam: It sounds cool though. Yeah. Jacob Luetzow: And it was like super easy to create. Adam: Yeah, that's that's awesome and that's like all peop a lot of people need. Jacob Luetzow: Yeah. And I still have a bunch of ⁓ domains like hosted on it. I just don't build websites for people anymore. Adam: How did like, ⁓ how did the how was the experience doing all that? Jacob Luetzow: I I think I just like hate sales so much. And like I hate just like talking to people for like this is gonna sound I I don't mean this to sound like I think I'm worth more than this. I just like hate spending my time for like a thousand dollars. Like I hate jumping on like a private personal call and then like most people like need hand holding and I I don't know, I don't have the patience. I like like high touch or low touch. Adam: Yeah. Jacob Luetzow: Customers where I just build a product so good they want to use it. Adam: Yeah. The thing is it's like that's how it is if I s if I do cold outreach to individual business owners, I have to do a call with them and I'm kinda like if they need anything like I have to help them and drop what I'm doing. And it's just not more scalable than a general tech salary. So it's not that motivating. Jacob Luetzow: Yeah. Right, exactly. Well, like and I think like that like my partner program where I'm reaching out to like other professionals, I think, yeah, it's probably worth a phone call with them, right? 'Cause they're actually gonna push my product to their clients. And I think that's worthwhile. But it's like at it's still tiring, like it's the last thing I wanna do, but it's like something I have to do 'cause I Yeah. Adam: Yeah. Yeah. And that stuff can lead to way bigger like that stuff's exciting. Jacob Luetzow: Yeah. And then like I also like I've gotten some good ideas too where ⁓ like a wealth management guy reached out. He's like, hey, have you ever thought of like allowing which I couldn't do this before because it was zero knowledge, so I kind of wrote it off, but now I could. He's like, have you ever thought about like having like roles? So like you could have like a wealth management or attorney set up dead man switches and put the files on, but they Only your client can reset the the triggers and edit it after it's been created. So it can kinda like kind of like your agency angle, but you let like someone else manage the assets in their vault. So I don't know. Adam: Yeah. That's interesting. Jacob Luetzow: No. But then also like how many different features can I think about to distract me instead of just getting more users? Adam: Yeah. Jacob Luetzow: I don't know. The world, the world and businesses. But yeah, all right. So back to yours. You push, you're pushing, you're focusing on agencies mostly because what? How how are how are you charging an agency like per client or how's that work? Adam: Yeah. So ⁓ I came up with like a tiered pricing, so like if it's one to if it's one site it's what you see on the website. If it's one to five it's cheaper. If it's five to twenty, it's cheaper, and so on. Jacob Luetzow: Okay. And then they're just slowly like can they start at a lower rate and then you auto scale ⁓ if they add people or is it like a Adam: Yeah, exactly. Just like once they hit the next tier, then those sites are all cheaper. Jacob Luetzow: Okay. Okay. Nice. So when you're looking for agencies, how are you finding how are you finding these people? Adam: Yeah, so this is where I use Glaud Code a lot. and I've tried I've like done a bunch of different contact scraping and tried different lists. But for the agencies specifically, you can use tools. ⁓ there's Apollo IO if you've heard of that. There's Hunter, so they're just like contact pools and you can sort by ⁓ Jacob Luetzow: Yeah. Adam: Company size, so I'll do like two to ten people work there. So it's a smaller agency. You can sort by the keyword of the company, so I'll do SEO should be in their description, or like web design. ⁓ that that's pretty much it. Jacob Luetzow: Okay. Dude, and if you do that, like if you go on like Google Maps and type that in, you probably will have like ten companies pop up in five square miles. If not more. Adam: Yeah, yeah, exactly. There's should be a lot there, so the TAM is big, I just gotta sell to ⁓ Jacob Luetzow: There you go. I'm curious, w shops like that, there's like like they're web developers or designers, right? They're slightly technical people. Do you find like that's a hard a hard sell? Or is it are they more eager to buy tools than be developers? Adam: ⁓ I think it depends. I like I can't really speak for them, but for me as a business owner, like if I mean unless it's really easy to code, ⁓ I'm probably just gonna pay for something because then I can focus on my backlog of stuff I already have. Like even at this point in my business, like it's not that big, but there's endless stuff to do. I feel like I'm working all the time. Jacob Luetzow: Yeah. Adam: ⁓ so if I can just pay for something and it's reliable and the quality's good, like I'll pay for it. And I I assume Jacob Luetzow: Yeah, I'm the same way. Like, cause you just want consistency and you want a good product. Like I'll buy if a tool works and it works well, I will pay for it. Adam: Yeah, exactly. Exactly. Yeah. And if if you like I think they can see the case studies that I have now. So it's like at least there's some proof that this thing actually works. Jacob Luetzow: Yeah, I think that's really cool. Dude, I feel like there's so many like niches you could go after, like contractors, like ⁓ trades, tradesmen, plumbers, electricians, so many things. Like I have a brother-in-law, he's in Australia, but he he paid a lot of money to have someone pump up his SEO, and all they were doing were writing blog posts and like Adam: Yeah. Jacob Luetzow: Not they they weren't doing a very good job either. So like I feel like like right there he would be an instant client. Adam: Yeah, that's the thing, like I try to just the goal is that this is a solution that's like pretty cheap and can do at least the content side of SEO and anything else I expand to. Because like so many SEO agencies just seem like scam artists. Jacob Luetzow: Dude, I mean, that's that's also like the problem. Like being technical and knowing how this stuff works, they'll promise things you can't prove. And I think that's the best part, right? Like they're SE, like a lot of like these companies will be like, you know, you can sign up for our top tier where we really focus on pumping your SEO and your Google search rankings and all these things for a thousand dollars a month. And like they're not doing anything different than. Adam: Yeah. Jacob Luetzow: Making sure, you know, you're a technical SEO and all your like lighthouse numbers are up, right? And then they're just adding new content to your blog every day so Google knows you're an active website. Adam: Yeah, exactly. Which like you know, you can almost with bloggy like you can see your technical SEO score. So you can fix that if you want to. You get the content. ⁓ the only other thing we don't automate is backlinks, but you gotta pay for those or like yeah. Jacob Luetzow: Backlinks are hard though. Like like you said, you're either paying for them or you're getting like a host blog post on a good website or like I backlink some of my buddies stuff on my like press page if it's like related to my product and that's it. Adam: Yeah. And it's hard to know which ones are even good. One experiment that I'm trying right now is I had Claude look at do like 30 different Google searches ⁓ for like AI SEO writer or like AI blog writer, B2B blog writer, all of those keywords that I'd want and find the first blog post that pops up and it's it got the emails of the author for each one. And it's just like asking them if we can get a backlink from them in a better way than that. Nicer way than that. But ⁓ I've gotten a few responses. A lot of them are just really expensive. But that is a tactic you can try. Jacob Luetzow: Yeah. Yeah, there you go. I'm curious now, all right. You support all these like direct integrations with like popular web frame like WordPress and things like that. Can you tie into an MCP and push content? Adam: ⁓ I should definitely add that. Yeah. I sh Jacob Luetzow: I would try it out. If if I could connect my MCP to bloggy, I would I would try you out and write some blog posts with it and see how I like it. Adam: Yeah. Yeah, I think for that, maybe I would do like a ⁓ more like pay per token or pay per use or something. 'Cause I feel like developers don't wanna subscribe, they just wanna write like a few and then or like have the keyword research part automated. Jacob Luetzow: Do you find it like very useful to have new content every day, or do you think that's overkill? Adam: I think that all the people that have gotten the best results did content every day, but like that can be like maybe a customer doesn't want to see that 'cause it looks more AI. ⁓ but honestly it does work the best from like a getting traffic perspective. Jacob Luetzow: Mm-hmm. Interesting. Yeah, because there is always that fine line, like, do these blog posts look too AI generated? Like even if they're div like delivering like good content, right? I think you still have a lot of like AI fatigue. Adam: Yeah, yeah, exactly. I think as long as the topics are different and it's not like the exact same topic thirty times next to each other, that helps a lot. Jacob Luetzow: Yeah, that makes sense. Cause like right now I do a new blog post every week and I'm like, do I need to step that up? Adam: Do you write them yourself? Jacob Luetzow: I have like Grok draft them and then I'll read them and like give feedback. Adam: Yeah. Yeah, I think ⁓ I mean that yeah, that's smart. Jacob Luetzow: Yeah. But like at the same time I don't want to like read and post more than once. See, it's like a lazy issue. So like if I could dial it in and just know the content is good, that would be awesome. It's just like though my bottleneck is reviewing code because I don't actually trust Claude to write good code. Adam: Yeah. Yeah. That's the thing right there that you said, it's laziness. I think it's just that you probably have a lot of priorities that are more important immediately. And that's why I think people will buy software and not vibe code it themselves. Jacob Luetzow: Yeah. I ag no, I agree with that completely. Cause like everyone, like you'll have your doom and gloom or your hob your positive people. There's always like someone wants consistent, like it's not like people act like deploying and having like something up in production is like the simplest thing ever. And like, yeah, with like past systems, it is pretty easy, but you're still spending money and then you have to make sure it's like reliable, it's up and it's usable and Adam: Yeah. Jacob Luetzow: Are there any open CVEs on it, like DF security vulnerabilities? Cause the second anything's open to the internet, you have a lot of things to worry about. Adam: Yeah, yeah, exactly. Jacob Luetzow: Like honestly, that like I come up with like new ideas all the time. And I like almost started I started building a new product and I was like, I'm being distracted. But like I was too lazy to get it deployed on a server because I was like, I don't want to go through all the DevOps stuff right now. But I like built it and I got it working on local hosts and it was really cool. I actually talked about it on my last podcast. I ⁓ wanted I was creating ⁓ I did create, I just don't think I'll Adam: Yeah. Jacob Luetzow: Take it much further than actually I don't it's not going to become a product I make money on, not anytime soon. But it was like an AI podcast producer. So it listens to you and I talking. It knows if I'm talking or if a guest is talking. It was transcribing on the fly. It would pop up. I had the whole app was a desktop overlay app. So it was like just transparent, but every like I had overlays of things. So it had like suggested questions or things I could say to add to the conversation. And then it also do fact checks. Like if something was said, it would fact check it and like show sources. It like it was really cool. But I'm not gonna take it to a product. Like I was just like, there'd be so much work into refining this to be perfect that I I don't want to take away my focus from kill switch, but I still think it could be a really cool product. Adam: Yeah. Jacob Luetzow: So anyone listening, you should build this. Adam: Yeah. Yeah. That I've always thought of not that exactly, but like like ⁓ when you listen to podcasts and people are like, Alright, pull that up. Like do it do a search. It could do that. Yeah, yeah. Jacob Luetzow: Yeah, yeah. That's like the Joe Rogan Jamie. That's what I wanted. It's like pull that up and then they'll like fact check it, right? And that's what I wanted. I was like, this is like the budget version of not paying for a podcast producer. Cause that's a salary, right? Adam: Yeah. That's that's a cool yeah, that's interesting. But yeah, there's endless stuff you could build. Jacob Luetzow: What's really cool is the way I built it, it ⁓ it would work with any video chat because it was a separate app using like, you know, audio output and input. It knew my input was always me. And then it knew output could be multiple guests. But what was cool is like I I was like testing it on meetings and things. So it like worked with Zoom. It worked with Google Hangouts or Google Meet, whatever it's called, and all these different platforms like it would work on here and I don't know. I think it was just it's kinda cool. So you could you could use it to record FaceTime calls if you wanted. Get the transcript and then you have the audio. Adam: Yeah, that is cool. Yeah. That's interesting. Jacob Luetzow: Yeah. And I I mean I was just like hacking together, like I I forget the product I was using for the transcripts. Let me see if I can find it. ⁓ deepgram. So I was using deep gram for my audio and it worked great. Like I just using their their API calls, I could just be like, and then also like this one's gonna have multiple guests and it would be like guest one. Guest and then like guest two would start talking and it would think it's guest one for a little while, and then all of a sudden it would figure it out that there's multiple people and it would go and correct all the tags next to the transcripts. It was pretty cool. Like, honestly, like worked really good. And I was just like kind of like duct taping tools together and just messing around. But like I think you could really narrow down like the way you prompt things and Adam: That is sweet, yeah. Jacob Luetzow: like add some guardrails and things to just really work the way you want it to. Especially like if I really wanted to like narrow it down, like make it for Elixir mentor. Like you could almost make the producer a a developer expert, like knows the resources and docs to check. And you could dial it all in. It could be really cool. Adam: Yeah. Yeah. Do you are you using it now? Okay. That's that's super cool though. Yeah. Jacob Luetzow: No, I'm not. I thought about it and I was like, I don't want one more thing to fuck around with. Yeah. Cause like right now, like it works really good, but it's almost like more distracting. And then I see bugs and things that I don't like. And then I get distracted from the conversation because I'm like, I gotta go fix that. Adam: Yeah, yeah, that's fair. Yeah. I mentioned I have a calorie counter app and like my mom uses it religiously. She has like a five hundred forty day streak or something. And I use it and some of my friends do, but not too many people do. ⁓ but like when there's a bug I still have to fix it even though like I really should not be focusing on it at this point. Jacob Luetzow: So I'm curious, ⁓ it does calories and you just talk to it about the food you ate. Does it track macros? Adam: Yeah, yeah. So you can set yeah, you can set goals. ⁓ and then there's like little macro bars that you can hit and you can connect it to Apple Health if you wanna do that and chat with like find like correlations between your sleep and your weight and your diet and all that. Jacob Luetzow: ⁓ that's awesome. Do you need to know like your serving size and stuff as you talk to it? Adam: Yeah, so like it can't it's not gonna be perfect because it's AI. It does use a food database, ⁓ but like the better you can describe the food you ate, like the quantity and like if it was fried or not, that's obviously gonna help. Jacob Luetzow: Mm-hmm. That's cool. I like that. Cause I feel like I always like the idea of tracking things. Like, and like I work out a lot and my biggest thing is not taking in enough calories, right? So, but I hate tracking it because I'm I'm lazy. So I but I like the fact you could just talk to it. I feel like you're more likely to record things if it's like frictionless to do it. Adam: Yeah, yeah, exactly. It's just it can be really tedious. That's that's why I built it. But working on it and fixing bugs is also tedious. Jacob Luetzow: I'll I like what do you have what's your user base? Like, do you have a decent amount of users? Adam: I did, so I actually got a lot of users. I probably got a hundred I I probably got like a hundred customers over time from SEO from Bloggy, but like most of them churned because the onboarding was bad or like it initially just didn't work that well. ⁓ and like I just I shouldn't be focusing on a calorie counter, it's too saturated, and bloggy just Jacob Luetzow: Damn. Adam: Make so much more per customer, so yeah. Jacob Luetzow: Yeah. See, that's the problem. You get distracted by all these things, right? Well, like I want to build a I want to get a 3D printer. And I was like, well, I need an excuse. I need an excuse to get a 3D printer. ⁓ and my buddy Alex just made a like an IoT nerves book. Nerves is like ⁓ the embedded version of Elixir on like IoT devices on like Raspberry Pis and things. And I was like, Well, I wanna build so what I want to build is ⁓ Adam: ⁓ yeah. Jacob Luetzow: We have chickens, right? And I like to let them out to like eat the bugs in the yard and it's like just good for their eggs. But they love to come up on our patio and they they poop all over it. And I wanna make like a motion sensing like squirt gun that will shoot at them if they cross the patio line. And I was like, Well, I need a 3D printer to make like the housing and do all this stuff and make it really pretty. So see, I come up with these excuses to build all these things. Adam: Yeah. I got I like tried building this robot ⁓ like a year ago, like a dandelion picker. That was like it's probably everyone tries to think of this kind of thing when you first buy a home. But so I got a printer for that and it is awesome. I don't use it a ton anymore, but it is awesome. Jacob Luetzow: Right. Yeah, see, I just need to get a printer. I'll find things to do for it, you know? My patio defender, it will be project one. Adam: ⁓ yeah, definitely. Yeah. There's there's a lot of good like household things like ⁓ like just like Tupperware lid containers, like that's one of the best things I've printed. Jacob Luetzow: That is pretty cool. Like just a missing part in the house, right? Just print it. I love it. All right. ⁓ this has been a lot of fun. ⁓ I really enjoyed this conversation. If they're where can people find you? Where can they find bloggy if they're interested? yeah, just spew out the links and they'll also be in the description, but Adam: Yeah. Exactly. Yeah, so I I talk probably too much on Twitter. My Twitter is ⁓ Adam Codes with a zero and a Z. ⁓ and then bloggy is bloggy. Jacob Luetzow: That's such a developer handle. Adam: Yeah, yeah, exactly. And then bloggy is b-la-w-g-y dot com. ⁓ but yeah, feel free to DM me on Twitter with any marketing questions or anything like that. Jacob Luetzow: And definitely follow Adam on Twitter. I I I enjoy your content. So I like file like I like finding people like I feel like you make really good content for like technical founders. Like I feel like it's educational. Adam: I appreciate that. Jacob Luetzow: No. But yeah, this is this has been a lot of fun and ⁓ thanks for joining me. And if you ever want to come back on and talk more marketing or whatever whatever else you've learned, just reach out. Adam: Awesome. Sounds good. Jacob Luetzow: All right, guys. Thanks for ⁓ tuning in and I'll see you guys on Friday.