Welcome to the panel where two bootstrap founders talk about building a best
Brian:Talk about choosing airline flight.
Justin:Talk about is that I say talk about? Discuss?
Brian:I don't even know.
Justin:I don't even know. It doesn't matter. Welcome to the panel where two Bootstrap founders discuss building a better business and a better life. I'm Justin Jackson, the cofounder of transistor.fm.
Brian:I'm Brian Casel. I am building Builder Methods.
Justin:And I think we got some good topics. Where where do we wanna start today?
Brian:What do you got? I feel like we've you know, we we always kind of veer into my stuff pretty quickly, but let's let's start with your your end of the thing.
Justin:So one thing I've been exploring is I had this post on Blue Sky where I said I gave an example of a bad experience with an AI chatbot. And I said, this is why we don't use AI chat in at on transition.
Brian:I saw you mentioned that in in the Slack. So what was what was the product? What was the question? What was your experience for for those, like, listening who maybe can't even see the screen?
Justin:Yeah. So I was we use crisp chat for all of our customer support. And I asked the folks at Crisp, is there an AI plugin that lets me query, query, query all of our customer conversations in order to get customer insights? For example, what adjectives do people use to describe Transistor when they sign up most often? So I'm I'm looking for like, I've got this whole trove of interesting customer information in live chat.
Justin:And I'm like, I wanna use AI to, like, bring up interesting you know, what are the questions that people ask when they sign up? How do people describe transistor when they sign up? Why are they here? You know, those kinds of things. So that's what I was looking for.
Justin:And then the their bot, their AI chatbot, confidently said, yes. We offer AI powered tools that can help you analyze customer conversations and gain insights, specifically the AI data hub. And here are some examples. And I was like, okay. And then before I to their credit, very quickly, a human operator jumped in maybe five minutes after this and said, please disregard the AI's answer.
Justin:At the moment, we don't have this feature.
Brian:Uh-huh. And and did you did did they do that proactively, or did you say, like, human? Like, talk to a human?
Justin:No. They did that proactively. So so somehow, you know, they're keeping an eye on this, you know, this stuff. But so this fits perfectly within my narrative, you know, transistor. We just released this new feature page that is all about customer service and kind of loudly proclaiming that we are all of our support is done by real humans that care.
Brian:I I'm I'm kinda torn on this on a couple different levels. One one thing, just an observation of like first of all, I'm definitely seeing a lot more of it across all SaaS tools, especially the big ones that obviously have a lot of volume to deal with in their support inboxes. Like, I definitely understand the case for it at Mhmm. At the at the like, the the large scale SaaS level where they clearly, they have just so much volume. They have to they have to use that at this point.
Justin:Yes.
Brian:Well, okay. That that could be arguable. Right? One thing that I don't like is when they just give them like a human name when it's clearly a bot. Yep.
Brian:And, like, you know, they might have a little icon that makes it look indicate that it's a bot, but they're just like, hey, it's Sam, your support agent. Like, wait a minute. Are you real? Or, you know, I I don't like that. Just just be totally upfront that it is a because I the other thing that I'm noticing is that they are getting better.
Brian:They're not great, but they're getting better. So I I sent us a question to Cursor Support a couple weeks ago about something. That response came back to me via email. So it wasn't live chat, but it was an AI generated response. And it was not clear to me that it was AI.
Brian:I mean, like, they didn't even tell me that They were just like, It's from Sam, the support agent. Yeah. And I guess you could say it was so good that by the second or third email, that's when I realized it was AI. I probably should have realized sooner, but part of what made me realize it was that it gave me the wrong answer. So
Justin:You know, my my preference is human powered support. Think I think customer support is the most underrated feature of most products. It's it serves so many purposes. It's your sales team. It's your onboarding team.
Justin:It's your PR. It's it they are internal sales. Right? So they're taking people that are already there on the website, and they are helping get them signed up and using the product. And I still believe that.
Justin:But I also want to be challenging my own beliefs. And Jesse at Bento, Bento Now.
Brian:Yeah. Saw his response.
Justin:He said, well and I didn't realize he's a one person team. That is wild.
Brian:I've always been I've I've sort of followed his stuff with Bento from afar. I've never really used it, but I've always been impressed with his efficiency and how much he's shipping. I mean, look at everything he to everything. Yeah.
Justin:Email marketing, messaging, marketing automation, CRM, sales, email delivery. Email is tough. He's competing with Yeah. ConvertKit and everything else. So he said his was good.
Justin:And I I do like how he's done it. So right away, you can see these avatars here are he has multiple AI agents. So I'm gonna ask a question. So Bento chat. So let me say, can I use Bento for transactional email?
Brian:Yeah. And that's like a straightforward question that it should be able to handle.
Justin:With the Rails app.
Brian:Like, give try to give it like a curveball. You know?
Justin:Yeah. I I mean, I'm trying to I try to have as many as much human stuff as I can. So here's what I like about it. He identifies that this is an AI support agent. So this is Tanuki, the general support AI agent.
Justin:And it was fast. Yes. You can use Bento for transactional email. Works really well with Ruby on Rails. Offer a native Ruby gem that makes it easy to send both marketing transactional emails directly from your Rails application.
Justin:And then they have these links to their documentation. And so what Jesse has said is that a lot of this your implementation of this depends on the quality of your knowledge base and Yes. Experimenting. So he's experimenting with multiple AI agents in tandem and just tweaking them and testing them with different characteristics to see how they respond. Like one agent is better for product related queries.
Justin:Another one's better for technical related queries. So I thought this was interesting just to one, to challenge my own kind of ideological beliefs and to test something like this out. There was a few things where it was, like, okay. It wasn't, like, as good as a human, I think. But and he also has it so that if you type human, you can talk to a human right away.
Justin:I I don't know if that's, like, becoming more common. Like, people just expect to be able to type human and talk to the is that, like, a normal keyword for that stuff?
Brian:I mean, I that that just seems like an age old thing from, like, calling up customer support lines over the phone. Yeah. You know, I would just say, like, human. Human. Like, get me to a person.
Brian:This is ridiculous. You know? Yeah. I I think that that number one, AI tools and and models in general are getting better. But I do think that when it comes to customer support chatbots, my guess or my assumption is that it's it's as good as you program it to be.
Justin:Mhmm.
Brian:And I feel like if you use any of these, like, totally vanilla, like, out of the box AI features that come with the help desk software. And I haven't done a lot of testing with them, to be honest. But my guess is if you just try to like turn that feature on and and let it go wild, it's gonna it's gonna create a bad experience for your customers. Yeah. It's just not gonna get it right.
Brian:So my And like, I have been thinking about this a lot for the customer For the Clarity Flow customer support inbox. Yep. And I haven't done anything about it yet. I just haven't had the bandwidth, but it is one of those tools that's sort of like in a queue of products that I wanna build just for myself. It's probably like three or four down the line, but like, I definitely feel the pain.
Brian:See, like, I feel like I'm in a unique situation with Clarity Flow because I completely agree with you that it is definitely the Bootstrapper's advantage for the founder or one person on their very small team to be personally in touch with customers. I've I've I've seen firsthand so much benefit both on the sales side, the customer development side, and just keeping paying customers happy. Like, having having best in class, like, super responsive, super helpful support from humans is
Justin:Yes.
Brian:That's something that the big competitors cannot do.
Justin:Yes.
Brian:Right? Yeah. So we need to use that to our advantage. Now, for me, as everyone who follows these stories knows, like Clarity Flow for me is in a mode where I want it to be as automated and hands off for myself as possible.
Justin:Mhmm.
Brian:But I'm still offering that human support, which is mostly me. Sometimes with the help of my my developer, we're still offering like really great support, which which means it's like time consuming support for us. And so my thought is because again, like right now for me, like the value of continuing to hold Clarity Flow and sort of give it what it needs to maintain its and do its thing.
Justin:I
Brian:do wanna try to find ways to automate and offload at least the tier one level support. Yeah. And I feel like theoretically, I should be able to, and like we use HelpScout and I know that they offer some level of AI features, which we haven't turned on yet. But my guess is like, they're probably not great. And also we have a really good knowledge base, like like a 100 plus articles that explain a lot of stuff.
Brian:Yep. But it's not through Help Scout. We actually host our own on our website. So my thought theoretically is to, at some point in the near future, build some sort of custom tool that will be fully trained on our knowledge base and be trained on the multiple years of emails in our Help Scout inbox. Yeah.
Brian:And how things have been answered. And I think those two things alone should give it a really good context. And then I would have a system to program in just hard rules. Like if someone asks about this or mentions this keyword, we're saying this, and we're not saying that. Right?
Brian:Yeah. And would just be a growing thing over time. And then I'm thinking through it like there should be a system for me to Well, there's also different levels. You could have an AI just draft your answer and not actually send it. And that could save a lot of time.
Brian:Yep. Right? And that could be like an interface where it's like customer sends in a question, AI drafts the complete answer, sends it to me, and all I need to do is click a thumbs up and it just sends. Or like a thumbs down and I fix something, and then maybe that fix gets baked into the training so that next time it knows. I feel like this kind of like logic needs to be worked out.
Brian:And Yeah. But I have an idea in my head for like how how you can and I'm sure Jesse has has thought through this sort of thing with with Bento. Right? Like like the and I I could just see from my day to day support load where it's like, okay, 30% of the time somebody is reporting a bug in the software. And so in that case, it needs to tag my developer who then replicates and fixes the bug.
Brian:Yeah. Right? In other cases, it's a feature request. They're not asking for how to use something, like, they wish it would in a certain way. We should tag that and respond to it in a certain way and maybe file it in linear.
Brian:And then there's a whole lot of questions that are just like, how do I use this feature? Or how do I For us, we have a lot of like, do I build the thing that I'm trying to build? I know Clarity Flow can do it. I just don't know how to put features all the together. Those are the most time consuming support requests for us.
Brian:Because we're essentially acting as like consultants for our customers. And it's hard. Like we I haven't figured out a way We've tried offering paid support tiers, which we've sold some of those and some customers are willing to pay. But even then it's like, I don't even wanna deliver that that service, you know, even if they're paying for it.
Justin:I think Jesse has done a really good job of thinking this through. I I like how the even I just tried to I opened up a new window. One thing is that it's very context aware. So if you're in a certain context and then you ask it a completely different question, it it doesn't do as well. But right now, like this one here, I asked it a a brand new question, which was due to recent federal ADA regulations, we are conducting an ADA compliance review for all web based platforms.
Justin:Could you please confirm whether your product is WCAG compliant? And so that to me is like, okay. This question is gonna gonna mess up the AI. It says, I wanna make sure you have the most accurate up to date information regarding Bento's ADA compliance. Currently, our document doesn't mention this.
Justin:So would you like me to connect you with our compliance and security team for a more detailed response? So to me, that's kind of the ideal interaction. It's like we we don't have that. They are linking to their SOC two compliance thing, article. But asking, hey, do you want me to connect you with a customer, a real human?
Justin:In this case, that seems like kind of the ideal interaction. The one thing that worries me is just even like having AI draft your responses. I think the danger there is what makes great customer service great is when it's human, where it doesn't sound like a generalized corporate speak.
Brian:Yeah. And There there is something about like the instant AI written response that makes me just not even wanna read it, even if it's totally accurate.
Justin:I just
Brian:don't wanna read it. Yeah. Actually, other thing that my approach to doing customer support on Clarity Flow now, cause like I'm consciously trying to minimize the number of minutes that I spend on a ticket without offering terrible support. Almost every support doc in our knowledge base has a video recorded usually by me. Yeah.
Brian:So our docs, they do have the text, but they're very video centric. So, so many of my responses to customers, instead of like me, like totally consulting with them and like giving them a whole game plan, like custom built just for you, I say like, Hey, here's the link to our Clarity Flow Commerce guide. It's got a ten minute video that you can watch and learn everything you need to know about how this feature works. And that should give you the roadmap to build what you wanna build. Yeah.
Brian:And like so many of my answers are just like, asked about Cloudera Full Commerce, here's the link to our video on that. You asked about our appointment setting, here's the link to our video on that. And so then when they watch it, they're still getting my voice walking them through how to use the software. So theoretically, an AI bot should be able to point people to a human recorded video. Like, the bot doesn't have to draft an answer.
Brian:But if they're asking about this thing, it's like, you know what? What you asked about is definitely in this video. Go watch it.
Justin:Yeah. Yeah. I'm so I'm gonna keep thinking about all this. I I it was I think it's a good practice in general to kinda even these things that I believe that are part of my value system that I believe strongly, I appreciate it. In this case, Jesse at Bento said, hey, I think I've got a pretty good one, and I think you should check it out.
Justin:And it was enough for me to go, yeah, that I should I should think again about this because maybe there is a use case for this. Right? Like, maybe our customer success people could rely on this more often and then invest even more in personalized Zoom demos. Invest even more in having, you know, recording personalized videos for high value customers. Right.
Justin:So there Yeah.
Brian:Like the that that the AI should be able to filter those. Right? Like tier one, it should be able to handle those. And then it should instantly recognize when something is a tier two or if it's a high value customer or something like that. The AI should route that to your people, right?
Justin:Yeah. Yeah.
Brian:And also I think that the low hanging fruit or the way to step into it would be to not have the AI be customer facing, but just use it to draft answers. And that right there probably removes so many minutes off of the support reps.
Justin:For training.
Brian:Yeah.
Justin:Yeah. This would be a great tutorial project, I think. Like, build an AI support agent that works in Rails is because I I believe that Jesse's is custom, And I think you're right. I think for this to be
Brian:From what I understand with Jesse's, it's his customers are using Bento for customer support.
Justin:Yes. Because the oh, yeah. He offers this as a
Brian:Like they offer it as a feature to for Bento customers to use, I believe. Yeah. If I understand that correctly.
Justin:I think that's right. Yeah. Yeah. It's wild. Anyway, that's what I got.
Justin:What what what about you on your side? What do you wanna chat about?
Brian:Let's see. So, you know, I'm I'm focused on builder methods. I I've been on a string of pretty good YouTube videos that have done really well, and then my most recent one was a flop.
Justin:Okay.
Brian:That's kind of fun.
Justin:Let's talk about it.
Brian:Again, if you look at my channel, the last, like, five videos, I think, five, I I think have have been all focused on this direction of building with AI. Like, clearly, I I pivoted about five videos ago. Mhmm. Yep. And then this most recent one, which is called cursor keyboard shortcuts for power users.
Brian:Yep. That's the one that I published three or four days ago. What is it? It's like four fifty views or something like that.
Justin:Yep.
Brian:So, you know, that is you know, YouTube I I really like YouTube's analytics. The It tells you It gives you a ranking, like a top 10 ranking for every video to see how this video is performing against your most popular videos ever. So you could see like at twelve hours in, twelve hours after the publish time, your video is on pace to be doing better than even your best video or second best or third best. Right?
Justin:Yes.
Brian:And this one, for most of its lifetime, was in fourth or fifth place.
Justin:Yeah. People said apparently, it's like that you get that one analytics where it's like, this video is performing, you know, way better in the first hour than your last video or something. I can't remember what Yeah.
Brian:I I use it on the mobile app. There's a there's a special there's a YouTube studio mobile app for for the creators to see to check their stats. Yeah. So that's usually what I'm looking at. But so this one okay.
Brian:So, like, what happened was, like, even when I was creating it, I knew that this there's a good chance that this one won't won't do as well as the other ones.
Justin:Okay.
Brian:And there's a few reasons for that. Number one is, like, Claude code is so popular right now. And my last two videos were about Claude code, and they did really well. This one I chose to do about cursor. I still use cursor every single day with Cloud Code.
Brian:But so, you know, cursor itself has sort of cooled off in terms of a topic for content. There's there's that. But the other thing is that, like, it's just about keyboard shortcuts, about my workflows. And there's I think there's some pretty good, like, hacks and stuff that I share that are kinda unique, but it's not like a burning problem
Justin:Yeah.
Brian:That that people are talking about, that a lot of that are people are gonna really resonate with, like, right away. Mhmm. One thing that I hope for that video is that, like, it this one might do fairly well with search. Like, people searching for cursor keyboard shortcuts might find this over time. You know, the other thing that happened was, like, last week when I recorded it, I I knew that I'm working on this new thing called Agent OS, which which I'm in the process of launching, like, right now.
Brian:Okay. Can I show a
Justin:screenshot of that, or is that still
Brian:on your left? Okay. You could you could show it at it's It is now live at buildermethods.com/agentos. I can get more into what that is in a minute, but just from the YouTube channel standpoint, last week, I wasn't ready to release a big video about Agent OS. It was still in the works.
Brian:So I was like, all right, I know that I'm at least two weeks out from having something ready for that. So I need another video to do in the meantime. And I need it to be quick and easy for me to record. Yeah. And I was like something that doesn't require me to like create a whole complex demo and and write a big script.
Brian:So I was like, alright, cursor keyboard shortcuts. I could just I can prepare that in a couple hours, you know? And that's that's what I did. So I sort of knew that it was gonna be like a average below average Yeah. Video.
Brian:But, you know,
Justin:On it's the other hand, you never know. This is the thing is I've had so many times where I've worked for days on a blog post, for example. And I'm like, this is gonna be it. This one's gonna be insane. It's gonna do really well.
Justin:And then I release it, and it doesn't do well. And then sometimes I'll just like write down a random thought and publish it in five minutes, and that one does really well.
Brian:It just blows up. Yeah.
Justin:You can know some of it ahead of time, but I think did you feel quite let down? Like, did you feel disappointed? Or
Brian:It's it's weird. Looks like again, like, I knew I knew going into it, like, once I hit publish, I wasn't very confident that this like, I definitely wasn't thinking, like, oh, this is gonna be a banger. You know? Mhmm. So it's like I I was, like, bracing for the fact that this was gonna be a a low performing video, and then it was a low performing video, and I was and I just felt like shit about it.
Brian:Yeah. You know, like like, it's just, like, it still it still burns. Yeah. You know? And and then and then the the irrational waterfall of of thoughts.
Brian:Like, oh, this is it. I, like, the the first few videos were a fluke. Like, this whole business is not gonna work. Like, this YouTube channel is never gonna take off. Like, it's doomed forever.
Brian:Like, which, of course, I know is all nonsense. But
Justin:This is Yeah. I mean, this is the hard thing about all of these platforms, I think, is, you know, why did I do well? And then how come this one didn't do well? And Yeah. It can feel yeah.
Justin:It can feel emotional is a good word for it.
Brian:Just The the one that I'm writing today, which I hope to record tomorrow and hope to release next week, which which will be about Agent OS Mhmm. Or or really, it's about this this concept that I'm thinking lot about called spec driven development. I think that's gonna be like a new term or something like it. Okay. I think it's a big concept.
Brian:I think it's pretty interesting and exciting and different and unique. And I'm presenting this new free open source thing that I created called Agent OS. So I I I I'm I'm feeling I like, I don't know how this one is gonna play when I release it. It it might it also might flop because it's so different. This one that I'm writing feels like I have something to say.
Brian:Yeah. Like, have something that I feel is pretty important Yeah. To say on this one. Where or it's clearly that one. I was just like, what can I record so that I so that I publish something this week?
Justin:Yeah. Is a danger in hopping on a trend or a wave. The the danger in that is you just don't know how long the wave's gonna last and whether you go and then it increases competition for that wave. And, you know, there are some valid questions around that. But I think clearly right now, like in software development, we're going be grappling with how do we do this for many years.
Justin:Like
Brian:Yeah.
Justin:It it the big companies haven't even really dipped their toe in. I know at some companies like Spotify, they're pushing it AI AI development quite heavily. But there's older companies that are still way Yeah.
Brian:There's older companies. I I think but but definitely, like, all I feel like all professional software developers are absolutely aware of it. And they're and, like, they're all moving to using cursor or and or Claude code. There are some that are holding out, I guess, for whatever reason. But, like, it's, like, it's it's and even even if you're just still using, like, Versus Code, then then you're probably using Copilot with it.
Brian:You know?
Justin:How important is it for you to publish every week? Like, did you did you think about just not publishing at all? Or
Brian:Thought about it. And and I've I've gone some weeks without publishing just because I couldn't fit it in. It is very difficult. Like, because and that's another big thing that I'm working on right now is like, how can I make my production process, especially the editing process faster? How can I and I think I'm getting better with every video?
Brian:I'm gaining little bits of efficiencies. Like, I'm I'm creating a bank of of b roll that I can use again, and and I'm I'm getting faster with the editing tools. Like but still, it takes me I right now, it looks like to make a really good video, it takes me at least a full day just to write the script and get the concept and get the get the pacing and get the message right. That's a whole day of just writing. Yeah.
Brian:And that's that's literally what I'm working on today. And and then tomorrow, if I finish the script today, then tomorrow morning, I'll spend the first half of the day recording it, like, probably the whole morning until lunch. And then and then I'll start editing it. But I lately, I've been editing, like, through the weekend And then and then to maybe publish it on Monday or Tuesday of next week. Yeah.
Brian:You know?
Justin:I mean, considering it's so much work, maybe that is it's last week, we talked about this, like, having alignment between what your audience wants and then what you're interested in, passionate about, curious about. And, yes, sometimes when it's like, it's like, I'm I'm just kind of interested in this, and my audience is just kind of interested in this. Maybe it's just not worth doing a video on that stuff.
Brian:But I I think that I'm going to figure things out on the efficiency side. At least I I just assume that I will over time. Yeah. It might might take me a while, but I'll I'll I'll figure that part out. I'm I'm more interested in, like, the the strategy
Justin:Yeah.
Brian:For the business and and the business model.
Justin:Yeah.
Brian:Right? And that's that's something that I'm starting to think about. It's it's an interesting thing because it's like being being the creator and and creating stuff on YouTube is such an a crucial part of this business. Yeah. Like, no matter what the products end up being, like, I'm gonna have to really embrace this role of, like, I create content on YouTube.
Brian:Mhmm. That's primarily what I'm doing. Right? So what does that mean for the business model? Right?
Brian:And you pointed you pointed me to this interview with Colin and Samir who are these big YouTube growth influencers. Right?
Justin:Mhmm. That was such a good that was Jay Klaus on Jay Klaus' Creator Science Podcast.
Brian:Yeah. Yeah. And I I've been aware of those guys, but I haven't really followed them too much. Now I I I started getting into them more. So I I follow them, and and I listened to that interview.
Brian:And it just started getting my getting the wheels turning a little bit. And then I'm thinking about, like, you know, this week with the cursor video that didn't do well, and I'm feeling the pressure of wanting to create really, really good thought provoking content, content that actually solves problems.
Justin:Mhmm.
Brian:It and I am getting inbound requests for sponsorships on the channel now, like, pretty pretty regularly.
Justin:Oh, yeah. They talk about sponsorships in that interview quite a bit. Like, that's one of the one of the things that gets brought up.
Brian:Yeah. Like, for most YouTubers, like, most people who consider themselves like a creator, I think for most people, they are thinking in terms of like, I want a big channel so that I can make money with sponsorship.
Justin:Mhmm.
Brian:And then there's a slightly smaller group who maybe think more like I'm thinking, which is like, I wanna grow the audience so that I can sell products.
Justin:Mhmm.
Brian:Right? And my assumption always has since the beginning has been like, grow the channel, grow the distribution so that that audience can get into my email list, and then I can sell products to them. Maybe courses and training and other products that they can Like, that's the way That's the path to a big business. And I think that's probably still true for me. Mhmm.
Brian:And so YouTube AdSense that I'm I'm just never interested in that. Like, I I could turn it on now. I just haven't yet.
Justin:Well, in the interview, Colin and Smear say they've always treated that. They've never budgeted based on that number. So they got their
Brian:The YouTube ad.
Justin:That's right. They got their first YouTube check for $500. They just said, that's always gravy. That's always
Brian:That's just bonus.
Justin:That's just bonus. But they
Brian:But the real money in sponsorships is like brands coming to you, and you have a private deal where in the middle of your video, you spend twenty seconds talking about their product. Right? Yes. And can once you have a significant audience, you can charge thousands for that per video. Yes.
Brian:I have those companies. Some of them are like pretty well known companies like in my inbox already. Even at my low level right now. Yeah. And I haven't started doing those.
Brian:My assumption has always been that at some point I will. Yep. But that won't be the primary business model. That'll either just be a bridge between here and the ultimate business model for Builder Methods, or it might just be a secondary revenue stream, the first one being a product line. Mhmm.
Brian:But I'm also just trying to open my sometimes I just like to ask the question, like, not that I actually agree with this, but like, what if the whole business actually was sponsorship?
Justin:Mhmm. I like this. Challenging your deep this is
Brian:the Challenging your thinking. Because I've always been a person who's just not very interested in a sponsorship driven business. Like, it's just that's just never really been interested interesting to me. But if I think about it, there are there could actually be some unique benefits to that approach in in this particular space. Like, because if my if I go with my first assumption that, like, the whole business is gonna be built on a few big courses and maybe one flagship course.
Brian:Mhmm. Right? Then that means most of my videos or a lot of my videos have to point people to buy those courses. Yeah. They don't have to be super self promotional, but they have to be on top.
Brian:So if I I will likely create a Claude code course in the very near future. Yeah. That's probably gonna be the first course that I create. Which means a lot of my content has to be about Cloud Code, has to be maybe beginner level that leads people into Cloud Code. I could veer off that a little bit, but at the end of the day, if I wanna sell courses, I gotta be pointing people to that.
Brian:Yeah. And maybe I'll have another cursor course. Maybe I'll have a different course about this or that. That could get tiring from a creative standpoint.
Justin:Yeah.
Brian:Right? There's also the the thing where it's like, you know, the tension between like, okay, I I want people to find value and pay for the course. So maybe I don't wanna give away all the goods in the free YouTube content. Yeah. Right?
Brian:So there's always that sort of tension. I know that you can make the argument, like, just give away just be as valuable as possible. They'll still wanna come for the paid stuff. But, you know, there is some tension there. There's also this weird tension that I felt in the past when I sold, like, the productized course where it was like, if I'm selling a course on this concept, then I own that concept, or I own that insight or that or that way of teaching things.
Justin:Mhmm.
Brian:So I always have to preach my my way my of doing things because that's what I'm selling in my course.
Justin:Yeah.
Brian:Right? Like, your whole business model is built on selling courses, then that sort of points everything you do in that direction.
Justin:Yes.
Brian:And that that can feel I think in in those ways, it could end up feeling kinda limiting.
Justin:I mean, there are people I mean, Aaron Francis, I think, is is wrestling with this right now. And I think he's come up with a combination of he's gonna do the nice thing about you and Aaron, and Wes Boss comes to mind as well, is your audience is software developers, CEOs, CTOs, founders. And there's these are customers that a lot of companies are trying to attract, and they're willing to pay big money. Whereas I think
Brian:You mean for a big money for sponsorships or for training? Okay. Yeah.
Justin:I think the the challenge a lot of other YouTubers fall into is if if it's if it's entertainment or, you know, if it's more of a consumer audience, then, you know, consumer products have all sorts of there's, you know, millions of influencers they could sponsor. And that's where you get the emails like, hey, can we do an integration for $300? Which is, you know, a bit more tough. But you've got this like high value audience, which should mean eventually I mean, we've got Levi in the chat here, but I'm assuming like a pretty good brand deal is like $5 per video. And it, you know, goes up
Brian:to $10. 15. What the rates are. You know? So but like then like the question is like, okay, if I'm opening my mind more to like, don't just think of sponsorships as the side dish, but maybe that's the main dish.
Brian:Then like, if you think about it, like it does offer in a way a lot of freedom on the creative side of the channel. Like number one, my job is just to create the most valuable, insightful content I could possibly produce. I would never think twice about holding back because my best stuff is behind a paywall in a course. Like, the goal in that case is just to make just to be the best YouTuber talking about this topic. Right?
Brian:Yeah. And that should be my goal with every video that I create. Like, if you wanna win at YouTube, that's sort of the mindset you sort of have to have, I think. Yeah. And then the other thing is just topic selection.
Brian:Right? I don't have to always hammer Claude Code or whatever the hot topic is or whatever thing I happen to sell a course on. I could just cover whatever topic I think is interesting because I have an audience that should care about what I think is important. Mhmm. It doesn't matter doesn't matter where we go with that.
Brian:Right? Yeah. As long as it stays within the realm of, like, building with AI. Right? Yeah.
Brian:And and that, like, so so then I'm also thinking about, like, this idea of how because I I'm gonna need to start to do this soon, which is, like, not just come up with a new banger video every single week, but I also have to have some, like, recurring formats, right, that are just easier to produce but are still high value. Yeah. So the two that I wanna get running soon are, one is gonna be interviews, like talking to other people. Show me how you build. Show me how you work with AI.
Brian:Let's pull some insights from that and put it up on the channel. And and I I need to think through like the format of it's not just gonna be like a straight like podcast interview. I wanna cut it up and make it really high value for it for a YouTube audience, but like bring other people into the mix as like guests that that I talk to. Yeah. And that could just be a recurring theme.
Brian:I can just book out a schedule of interviews, and that'll be a stream of content that can be coming up in between my insight videos. Right?
Justin:Yeah.
Brian:And then the other one is, like, watch me build something. Right? Like, I have this whole list of product ideas that I wanna build myself for for my own use, or maybe it'll release as real software products. Like, these should be YouTube videos, maybe even series. Right?
Brian:A single product might be a series of three to five videos of the whole life cycle of the build. Right? And that could be something that people will tune into to see to see the next episode because they want to see how it evolves. Right?
Justin:Yeah. Just to pause. I mean, I think this is worth exploring for sure. The challenge with sponsorships, I think, is kind of multifaceted. One, it can be difficult finding sponsors consistently.
Justin:And you might get like one or two that I think Levi was the one telling me about this. But, you know, he's had some that paid him really well, and it was like, that was amazing. And then there's just like this dry period where it's like nobody's showing up.
Brian:Yeah. What I'm doing now is like every time someone sends me an email, I just tag them in my email with like potential sponsors. Yeah. So that, like if and when I need to drum up sponsored business, I have a list of people who've who've inquired.
Justin:Yeah. But my guess is a large portion of those are going to lowball you.
Brian:Yeah. They're gonna be low. Yeah.
Justin:It's also worth looking at what Jay Klaus has shared in the past where he makes his revenue. And sponsorships is not it's not insignificant. It's 15.6% of revenue. And memberships is his big 47%.
Brian:Yeah. I feel like Jay is more in line with what my my main vision for builder methods is. It's like, grow the audience somehow, and I think it's YouTube, and get them into a flagship program, then plus some smaller programs. Like that, Jay has this membership with the lab. Mhmm.
Brian:And then he has a couple of other paid courses that people can get into.
Justin:I believe Colin and Samir do this as well. I think they have Mhmm. YouTube starter school or startup school or something.
Brian:Yeah.
Justin:I I'm not sure if that's the majority of their revenue. I mean, I'm sponsorships are interesting to me right now because increasingly, I'm just seeing more and more companies invest their marketing dollars this way. We've talked about this on the show, which is traditional like Google Ads right now. That marketplace is so efficient that it's really hard to eke out any value on top. So you'll pay a really high cost per click, and it's super competitive.
Justin:And you're just lucky if you can make 15% on top of that in terms of profit. And so I'm seeing founders like Jesse at Bento, He's sponsoring everything. He's sponsoring the CSS homepage. He's sponsoring mostly technical podcasts. It just he's just trying to show up everywhere.
Brian:It does seem like the most value or it it does seem that like, I like, I could totally buy the argument that, like, sponsoring creators Mhmm. Is gonna be the marketing channel that's left after everything else becomes untenable.
Justin:Yeah. And it's an interesting one because one thing that Jay said in that interview was he's realized I mean, this makes it more difficult too, but he's realized that every sponsorship deal he takes is an endorsement.
Brian:That's that is what's tough about it. Like, I like, you would have to, like, really use and believe in the product. Mhmm. And the product can't really conflict with, like like, if if the video is about Claude Code, I can't I probably can't do a sponsorship that's like a competitor to Claude Code or
Justin:something
Brian:like that.
Justin:Right? I mean, I think the one version of this that's interesting to me is if you could get a big sponsor to just give you a retainer.
Brian:Yeah. Got a lot of them.
Justin:Yeah. And then it's like, we'll pay you $5 a month to make a video every month. As long as it's about Claude, we will
Brian:I I think that that is sort of the future of a lot of marketing, especially for SaaS, honestly. Like like, there's always gonna be different channels that you can do and be successful with. But SEO is not gonna be here forever. Ads are have have have become really difficult and expensive
Justin:Mhmm.
Brian:For many years now. Honestly, like listening to these interviews with with Colin and Samir, like the like, it does seem like this this is nothing new, but it it it tracks. Right? That, like, most companies can't just have a creator on staff to to create themselves a YouTube channel with hundreds of thousands of subscribers. So so their their next resort is to go sponsor creators.
Brian:Yep. Right? I I just think that that is gonna be an exploding marketing channel for most company.
Justin:Yeah. You know, we we saw this with, I think Wes Bos is just another incredible founder in this space. And he, you know, he's been doing this a long time. And I that's why, you know, early episodes were sponsored by FreshBooks, LogRocket, you know, and clearly, his high value audience of software developers and CTOs, people were willing to spend the money. And this worked so well that then they got purchased by Sentry.
Justin:So now Syntax, Wes Bosses podcast, is basically just a a giant sponsorship for Century. And Right. I'm assuming they're getting a lot of value out of it. They they just see it as a good channel to have, you know, Wes, and I can't remember his co host, you know, chatting about it every week.
Brian:I follow him, but is is he primarily sponsorships or does he he sells courses too, doesn't he?
Justin:Oh, this is I just think a small for him, I think this is a small piece of his pie because I think most of his He's
Brian:not in, like, JavaScript course.
Justin:Yeah. He's like the the he's got a course machine. You know? Yeah. Yeah.
Justin:So courses you can often tell what's the most important bucket. You know, we always talk about buckets here. Like, your most important revenue bucket by what you list first. So courses is first. Syntax is next.
Justin:So I'm assuming the revenue goes in that order.
Brian:Like, I'm still thinking of, like, courses is probably still the way I I would wanna go. But, like, I'm I am starting to, like, to open my mind to it a little bit more. And and also, like, I'm still it's sort of it's still sort of an open question to me. Like, is there really a market to sell courses anymore? Or is that gonna go away?
Brian:Yeah. With, yeah, with AI, but also just in general, like, people learn by doing, I would I would think, more more than ever.
Justin:But still think there's a lot of value. This is where Jay Klaus is going as well. A lot of value in community, in cultivating a community and having people pay to be a part of that. I'm even looking at Jim Hill has this this comment here on the live chat about making a course to deal with your emotions as a developer in relation, I think, to to AI coding and stuff. And
Brian:You know, it's fun. Like, really, the the video I'm working on for next week is sort of in that vein.
Justin:Yes. But this is a great place to put community. So generally, what I am seeing I this is actually not I don't think what Jim was talking about. But generally, in the market, what I'm seeing is that there's a lot of developers very anxious about AI and coding. And they might be curious about it.
Justin:They might be, you know, interested in it. But they are worried about it. And those are the kinds of things I would want to express in a private community. Like, hey, team. Right.
Justin:We're all here together. And we don't want get left behind. And we are also trying to you know, we have a job to do. We got to please our boss. We got to get these deliverables.
Justin:My boss is saying I have to use this. And there's the practical side, which is how do I you know, like, I was asking you the other day for an existing Rails project, should I use Cursor or Claude? But then there's the emotional side, which is how are y'all feeling? What like, here's a a place where I can go and express my frustration or my fears or my wins.
Brian:So that's I I definitely agree with you. Like, that would be the value for most paid courses or memberships or combo of both? It's like, yeah, of course, you can get all this information for free on YouTube or just by reading the docs on Cloud Code or whatever it is. Of of course, you could just go figure it out just like everyone else does. But the value of paying some amount is to learn alongside your peers, build stuff together, share notes, join a community.
Justin:But
Brian:like, it does raise the question again, like, what am I actually trying to build here and how do I wanna spend my time and resources as the founder of this?
Justin:Mhmm.
Brian:So just thinking like fundamental questions, like, again, like, sort of opening my mind to, like, if if it's a sponsorship driven business model, then the only thing that matters on my schedule every single week is to build stuff so that I can teach it
Justin:Yeah.
Brian:On YouTube. Right? And create really good content. Like, that's the only thing I ever really need to do. Mhmm.
Brian:If if it's a courses business, then in addition to creating good content to keep the top of the funnel going, then I would also need to create courses, which is very time consuming. Yeah. And if and if there's a community membership component, then I have to be active and be the community guy and and offer some other exclusive value to the members. Yeah. I don't know what that is.
Brian:And, like, that that piece, which I which I think is the most valuable thing, is probably the thing that I'm, like I'm just not really cut out. Like you are and like other people do a really good job with that. I like to participate in communities, but I'm not a big like rah rah, like lead the community kind of person. You know?
Justin:Yeah. My inclination I could be wrong, but my my inclination is that it's probably difficult to create a business that's just based on sponsorships because of all the variables we described and also just the risks. That market can fall out underneath you. You know, like, the stock market goes down and everybody you know, we saw this when everybody was cutting their developer relations budgets. All these DevRel people Yeah.
Justin:Who had been hired to be exclusive YouTube creators for all these software companies got fired in large numbers. Like, it was for a while there, just every company seemed to be getting rid of their their dev rel team.
Brian:I agree with you. I I do think that, like, at the end of the day, like a mix. Yeah. Because that that's the that's my current assumption is like it's it'll I I will be doing sponsorships at some point, but it won't be the main thing. And I will be doing courses, but I I it's it's just an open question for, like, how I I also think that there's risk in courses too.
Brian:Yeah. But I but I think that, like, it's probably a combination of both.
Justin:Yeah. And we can see, you know, if we look at it it can be helpful to look at the template that other people are using. And so Colin and Samir have Creator Startup. Jay Klaus has Creator Science. Wes Boss has all of his courses.
Justin:Aaron Francis has his courses. It feels like it always has to be a both and. It's like we do the sponsorships. And that's not to say maybe you shouldn't just focus on sponsorships right now while the market's hot, because the market is hot right now. And so maybe for the next six months, it's just like all I'm focusing on is getting that sponsorship revenue while it's while the window is open.
Brian:Yeah. Part of me is like, I I've I haven't taken any yet. My thinking is like, I still wanna grow the channel first and just focus on like, I'm still experimenting with every new video. It's an experiment to see like, what is really hitting? How can I figure out the formula for producing really good content?
Brian:How do I have more bangers than flops on a more consistent basis? Right? Like, I'm still figuring all that out. Yeah. I do wanna get to revenue as soon as possible.
Brian:I think I got the domain. There's nothing there yet, but I got claudcodecourse.com. That was available. But yeah, Agent OS is Yeah. It is a legit product, but it is free and it's open source.
Brian:And I'm pretty excited about it because I think this will be the I'm I'm gonna be using it heavily, but it it is the basis for a lot of the stuff that I'm gonna be
Justin:Tell tell us about actually, first, before we go to Agent OS, I just wanna say one more thing.
Brian:We could flip back to you, whatever you
Justin:No. That's fine. I think let's let's end on Agent OS. But with regards to community, I think there are ways you could do it in like, you are really good at process. And even thinking about what you did with productized services in the past, I think you could either hire somebody to be a community community moderator.
Justin:Or, you know, in Mega Maker, the community I've run forever, we've just had volunteer moderators. And I think paying somebody to do it is probably best. But you could if if there was enough there, you could just have somebody who's it's like, Brian is doing this stuff. He's making YouTube videos. He's doing it.
Justin:But, you know, Janet is her whole job is just to be onboarding people into the community, helping them, you know, craft conversations and and hosting calls and things like that. Like, is a world where you could do that if there was Yeah. Enough
Brian:That's the thing. Like, hardest thing, of course, about a community is like the starting point. Getting the very first members and keeping them engaged and really have that launch pad for it. I think there's probably two models that come to mind. Like you talk about, if it's more of a evergreen, like like you're you're participating in this thing all year long, maybe multiple years Mhmm.
Brian:That's probably a lower price point. Like and and it's a higher volume game. Right? And that's one way to go, especially if the audience can can grow to high numbers, then then you can have and and I see a lot of YouTubers in my space doing this. Like, their main thing is like a $50 a quarter or $50 a month.
Brian:Yeah. Like, just access to a private whatever
Justin:Yeah.
Brian:Discourse, whatever it is. Yep. And, like, that's one way to do it. But I again, like, I don't I don't maybe there's enough maybe eventually there would be enough volume to do something like that. And But then the other thought that I had was maybe the flagship product becomes a timed cohort based thing that I only do two or three times a year.
Brian:And basically a pop up community
Justin:for I like that idea. For
Brian:six weeks, we have a tight knit community. And then after that, yeah, maybe you can stay in touch, but the obligation is over. The event already happened. Yeah. And with the cohort based thing, that would be a higher price point.
Brian:So that could be a significant product. And part of the value is sort of the urgency. It's only happening now until we do it again six months from now.
Justin:Yeah. Yeah. I think those are really good ideas. I I I like the idea of pop up cohort based community because then it's like, you're not promising people this is gonna be around forever. Yeah.
Justin:And, you know, it can have there's is a benefit to being like, we're this little group. And you guys want to keep this going afterwards in your own group, like, can keep this whatever WhatsApp group open. But in the meantime, we're going to be kind of going through this program together. And when it's done, I'm out. I'm I'm gonna go back to to not being here.
Justin:But I could see because you're still there is a lot of value in being the organizer. Like, not everybody can attract an audience. Not everybody can seed something like that.
Brian:Yeah. Like, I I definitely enjoy it. And I've I I love interacting with with people in my audience and customers with my products. I I love that. But I have done it a few times.
Brian:Like, I did it with the productized course, and I've tried to do it with other products. And for me, it has always fizzled out pretty quickly because, because, A, I just didn't have the volume. I didn't have the number of people where there's just enough engagement. And then, B, like, I got disengaged, and then it just fizzled out. You know?
Brian:And that that's always my fear with anything membership community based.
Justin:Yeah. Yep. Alright. Let's talk about agent OS, buildermethods.com/agent0s. What what is this?
Brian:Okay. So I'm still trying still trying to dial in, like, the elevator pitch for it, but it's it is a system, an operating system that you can give to your coding agents. No matter which tools you use, whether you use call code or Cursor or I don't know if Windsurf is even still a thing anymore, but like, you know, any of them. So it's essentially a set of instructions and commands. So these can be cursor rules.
Brian:They can be claud dot m d memories. But there is a system where, essentially, it it's giving your agents even even more context than they would normally have. So like currently or up until now, most people just kind of write a really like just write prompt after prompt after prompt to try to get it to build this next idea or build this next feature. Then they're going in circles. It doesn't build the right thing.
Brian:It builds it in the wrong way. You gotta go back and fix things. It didn't build things the way you would build it. ClonCode might be very capable, but it still doesn't have your style or your thinking in on it. So this get this really provides like three things.
Brian:Right? Like the the main goals of it. And you can see this on the on the page that I that I set up for like, would you wanna use Agent OS? The first one is it has like complete context, not just the prompt. Right?
Brian:So at at a base level, you're setting up your standards. You're gonna have a folder on your system that has your coding style, your coding best practices, and your go to tech stack. So, like, for me, I'll have a tech stack file that that you know, it's like Rails, Stimulus, Tailwind.
Justin:Oh, yeah.
Brian:You know, I like to I like to use import maps. I like to use Google fonts. I like to use loose side icons. Like, I I always deploy this way. I always use that hosting provider.
Brian:My go to stack, right?
Justin:Man, this sounds like it'd be good even just for improving the way you onboard new developers.
Brian:Absolutely. And I'm glad you said that because this is the whole idea, to actually treat your coding agent like it's a new team member that you're onboarding into your company.
Justin:Yeah.
Brian:Right? So you're not just prompting it, You're giving it the context of, this is how we build here. Mhmm. Like, you just joined my company, so I have certain opinions on how like, for example, another file in my standards folder is best practices, like coding best practices. Right?
Brian:Like for me, I tend to prioritize writing code that is highly maintainable over performance. Like, yes, we want it to be performant, but it has to be easy to read, easy to maintain, easy to iterate on. Right? And then I also have like very strong opinions on like formatting. I have some weird opinions on like how I format TAL and CSS classes and how I do indenting and how I do this and that.
Brian:Like Yeah. I have examples in this file where, like, when you write code, I want it to look like this.
Justin:And is that all on the GitHub like, in the GitHub repo here?
Brian:Yes. But there's some sort of like generic standards that part of the part of the thing here is like you need to set your own personal standards. I I give you a template, but that's sort of like step one is you set your standards. Right? And then step two is the product level.
Brian:Right? Or the right? So like then Agent OS has the instructions built in so that when you are either planning out a new product or you're plugging it into your existing product, it's gonna create a folder inside your project called product. And inside that, it's got the mission statement. It's got your roadmap.
Brian:It's got your tech stack for this product, for this code base, and it has a log of all the decisions that have been made for this product. So we're we're not yet getting into building features. We're just setting up all the all this context. Who is this product for? What's the problem that it solves?
Brian:Who's the target customer?
Justin:Dude, I wanna do all of this just for regard even if I'm even if I'm not doing and to have a template for this is so helpful.
Brian:So the product folder in in your you have the mission MD. It it's got, again, like, the jobs to be done, target customer, the mission, the goal for the business. And then the roadmap, it is a checklist, but it's not like actual tasks yet. It's just your product roadmap, like high level, you know, maybe like eight big features that we plan to build over the next couple weeks. Okay.
Brian:So so we have our standards. Like, this is how we do things here. Then we have our product. This is this is the story of why this product exists and what we're building here.
Justin:Yeah.
Brian:And then third level is the specs. Right? This is where I get into spec driven development where it's like every new feature that you build or every improvement that you're making to your project, you're gonna have a spec. Well, there there's a specs folder, and inside that, you'll have a folder for like, okay, we're gonna we're gonna build the user authentication now. Right?
Brian:So inside that folder, which which has a date on it, then you have sort of like a PRD, like a spec document. Like here's the goal of the feature. Here's the job to be done. And then you have API specs and database schema and technical requirements. All of all of these are like individual MD files inside each individual task.
Brian:And then, of course, you have like tasks.md just for the just for that feature. So we're building user user auth. We're gonna have a tasks file with a checklist, maybe, like, three big pieces and then, like, six sub pieces under under that. Yeah. These are and then and and then, like, that is what the agent is gonna work on.
Brian:Right? So when it works on it, it's going to be able to reference back to your standards, how we build things here. It's gonna reference back to your product, what it is that we're building here. And it's gonna give you the checklist that like, we've seen we've seen this pattern come up now where, like, agents need to follow a checklist. And, you know, it starts with, like, writing a test, then building it, then making sure the test passed.
Brian:And I
Justin:I I just wanna pause you here to say this is awesome. This has that feel of it instantly ignites a spark in me because I start imagining myself using it or implementing it. I think you're really onto something here. And Zach was saying this in the chat as well.
Brian:I think so too. Like, I've I've been using it, like testing it out on on my own work. What's also really cool about this is that it really gets us away from writing extensive prompts.
Justin:Yeah.
Brian:Because right now, without this, I would have to write an extremely detailed prompt to overexplain, like, here's how I want you to build this thing, and here's sort of the back story on why you need to do this. And then let me point you to this and that file. But now with this installed, literally all I need to do is type in, what's next? And then it will automatically go to my roadmap and say, alright, well, next on your roadmap was to build the emailing feature. Should I go ahead and plan a spec for that?
Brian:And I'm like, yes, go. And then it goes ahead and writes the whole like it writes out the whole user stories and the task list. And then what I need to do is I just need to spend the time to review and correct the plan before it gets to work.
Justin:Yeah.
Brian:That's what we, as developers, need to start to shift into. We are now specs creators, but really the AI writes it for us to review. We dial it in. And then when it's good, we say go. Yes.
Brian:And and, like, that's how you get to a point where AI is, like, building not only what you want it to build, it's building the way that you would build it because it has all of your ways built into it. Because it's still a lot of work today to write these extensive prompts. And this this just removes that from the picture altogether. And really, the work now becomes like, go ahead and plan it for me. And now let me correct you.
Brian:And now now I approve you to go ahead. And that's it.
Justin:Yeah. You you painted the picture really well there. I what what's your like, does this need to be free? There's some people in chat saying, you know, crazy that you're not charging for this. You know, this seems to be like, even in chat, the reactions in chat are people are resonating with it.
Justin:So what does it need to be free? Is it what what's the game plan in terms of, like
Brian:I don't know. Like, I don't know. My thought is like, at first, I was like, I don't know. Maybe maybe this is a product. I built it like it's a real product.
Justin:Mhmm.
Brian:It even has these installer scripts. So it's like a one liner. You just paste it into terminal, and it'll, like, it'll put these instruction files on your system where they needed to go. So that that's pretty cool.
Justin:Yeah. I mean, I I'm kinda conflicted because it's like
Brian:Like, my thought the reason why I made it free is there's already a lot of people recommending and providing, like, template, like, cursor files. Mhmm. There a lot of them are just like, if you use a React stack, then here's a cursor file to to to do that. Or here's a long list of all these different cursor rules that you might wanna fit into your project. But to me, all of those are just like, they're not really a system.
Brian:They're just like a lot of words to throw onto your project that may or may not help. You know? Yeah. This is more like tailored to you and you it you need to shape it to how you want your agents to work for you. You know?
Justin:I mean, the benefit of it being free is it is very awesome. It provides a lot of value. If you get a lot of people using it and starring it and sharing it and, you know, that usage could translate into something paid on top. Like
Brian:That's my thought. Like, I and and then I was also thinking, like, alright. Well, maybe it'll be free, but maybe behind an email wall. Like, maybe I can turn this into an email lead magnet.
Justin:Mhmm.
Brian:I gotta fix the camera. You know? So but but even that too felt a little bit clunky. So I was just like, well, if I'm just gonna put it up on GitHub, let let me just make it totally free. I ended up putting the docs on the builder method's website.
Brian:Yep. So at least if you if you wanna get into the docs, you're gonna be coming to you're gonna be traffic on my website. And and then hopefully from there, you know, that that will get you into my orbit somehow.
Justin:Is this ready to go right now? Can people try this already?
Brian:Yeah. It's all it's all there. The the only thing that's missing is the video. I haven't I haven't posted the video yet, but the those those scripts, like, installation, like, you could install it right now.
Justin:Do I have to install this in a specific, like, project folder or this is just universal?
Brian:Universal. You copy that, you drop it into terminal.
Justin:Dude, this is really cool. I mean, it's the age old problem. You're hoping for a lot of people to use it and then for free and then to, you know, offer something on top. You know, there's ideas.
Brian:And this is totally new for me. I have never believe it or not, I've I've never really released an open source product. Yeah. Like, my my personal GitHub is blank. Like, I I don't have any plugins or anything like for free.
Brian:Like, I've just never done that. So I don't know what I'm doing here.
Justin:Well, I mean, I think the first thing is everyone needs to go. The link for this will be in the show notes. Go and star this on GitHub. I think stars is one way of getting some traction on GitHub. That's thing we need to start asking.
Justin:And Alright. Start using it.
Brian:Like I like, when people start posting, like, GitHub issues or PRs, like, don't know what to do about that. I I I this is a whole game that I'm like I don't know how to manage an open source project, but we'll see how this goes.
Justin:What I like about this is especially if people share it.
Brian:My hope for it this might be a total flop. Maybe it's too much for but my hope is that it's something that people talk about and refer to and either to point to the website or hopefully soon to point to the YouTube video that I'm gonna do about it, which I'll I'll publish next week.
Justin:I mean, obviously, the the way you present it is pretty important. I I think what one challenge with this is like like I these first lines on your landing page right now, they're good. But the magic I started to feel the magic and the spark when you got more into the details. And I was like, oh my gosh, like, I would use this just to improve, you know, our onboarding for developers or just for improving our the way we work as a team already. Like, that's already solving a problem.
Justin:And then this is has this added bonus of now we can figure out our AI driven development stuff. Yeah. So however you present this
Brian:Those docs, I've gone through so many iterations on them already. Like, just that section there, like, I went through like multiple iterations with the help of Claude helping me write stuff.
Justin:This is great. And I think it will lend itself to I'm hoping, and maybe you can even engineer this a little bit to the kind of sharing, like, this tweet from Adam Waddon, bullish on zed, feels like I'm using Sublime Text again, but with all the modern niceties that originally pulled me over to Versus Code and Cursor. So so people sharing stuff like this, it feels like Agent OS could have this kind of momentum. Like, people are just like, dude, I found this thing that really works for me now in this context, in this AI driven world. Like, this is what everyone should be using.
Brian:I I hope so. And and like also, think an argument for why it's free is it's intended to be your system for either you personally or for your organization. Adapt it like, it's meant to be Mhmm. It's not like so so, you know, there's these tools like Taskmaster and even this new one that that just came out, the the Amazon owned is it Kiri?
Justin:Oh, Kiro or something?
Brian:Kiro. Like, that's another Versus Code fork. Like, there's a lot of these tooling out there that have something like this sort of with specs built into the process. But all of that stuff is like super rigid. Like, you have to follow their specific interface for how it works.
Brian:And it and it's not like a system that you can really shape to your methodologies and your opinions on code and and all these different things. And like so so this is designed to be it's just a set of markdown files. You can use it with Clog code. You can use it with cursor. You can use it with Kiro.
Brian:You can use it with any of them. Yeah. And you can and it's not intent like, the instructions that are built into it, I think, are good to go. But you have those are open source too. So if you want it to execute tasks slightly differently from how I designed it, you can go in and tweak the instructions.
Brian:You know?
Justin:Yeah. And so what somebody mentioned Taskmaster and that they have some sort of business model with this, but I can't what is their business? Do you know their business model? I don't
Brian:know what their thing is. I from I don't I don't really know. I I played with it, but I think that they're planning to build some some upcoming products.
Justin:Okay.
Brian:I don't I don't know what it is. But I'm I'm definitely not trying to build a code editor. That's
Justin:Yeah. Yeah. Yeah. Yeah. Yeah.
Justin:Cool, dude. Well, I I think this is great. And this is one of the nice things about doing the show live is that we get to see everyone's reaction to it. So well done.
Brian:Yeah. And I I I think I think it'll be a really useful thing, but also for me as a teacher and as a trainer and as a coach, like Mhmm. A lot of that content is gonna point people to like, alright. Well, my way advice is to use something like Agent OS, you know.
Justin:Awesome. Well, if if anyone has ideas out there on what Brian should be doing that with this next and how he should be managing it as an open source project, let us know. I think this is good for this week. I think we we covered a lot of good stuff.
Brian:Yeah. What what do you got going on over the weekend and into next week? What are you what are you looking forward to?
Justin:What is going on this weekend? Got a friend's birthday. I'm getting ready to go on this this trip to go pick up my son so he can he's been at college. He's coming home from basically living in that city. And then we're gonna get him ready to go move to Toronto.
Justin:So that's gonna be a big
Brian:Oh, cool.
Justin:Big trip.
Brian:Moving out.
Justin:What about you? You do anything fun?
Brian:The kids are gonna spend the weekend at at at grandpa's house.
Justin:And Oh,
Brian:yeah. We're gonna love a date night probably. May maybe go see that f one movie.
Justin:Oh, yeah. Yeah. I wanna see that new Superman movie.
Brian:Yeah. I've heard I've heard mixed reviews on
Justin:I'm not
Brian:a big superhero guy, but I I know that one
Justin:seems Well, and I'm not a big Superman person, but I really liked Guardians of the Galaxy. So I and it's the same guy. Right? So, yeah, I wanna check it out.
Brian:Yeah. But I I've heard really good things about this f one movie. I'm not a I'm not an f one fan, but I heard that's I heard it's the kind of movie that you wanna see in the theaters.
Justin:Okay. Well, maybe I need to see two movies this weekend. Yep. Alright. Thanks, everyone.
Brian:Alright, folks.
Justin:Later. We'll see you later.
