- Platform: meet
- Source: Read.ai 01KRBT5S5A122EHZMVXPWAR724 (via Ben Ornstein’s account, imported 2026-08-06)
Transcript
Ben Ornstein: That’s going to be huge, I think, because that makes total sense as one of the biggest problems of the process. Michael Houck: I should have flagged it to you, because I have this, I think it’s like 250 messages or 300 messages at this point, like YouTube. They don’t want to start a new thread and have to reteach it YouTube. So I’ve done all the scripts in one thread. But I keep doing that with it. I’m like, no, give me angles first. I should have flagged that to you as emergent behavior that I was noticing. But it’s going to take time. Ben Ornstein: It makes total sense. And I think that will, that’s going to increase the quality like substantially, I think. Yeah, I think that it’s going to be a big one. Michael Houck: Okay, cool. So we wanted to use this time to talk about next steps for the agent team across different verticals. Excuse me. And I think also consider the role of paperclip, that thing I sent you, that open source project to like, let you have the hierarchy of an agent team and like how, what they’re supposed to do, their responsibilities, their org chart. Right. I think we’re going to want to think about the role of that because if we have agents across, you know, audience building, geo creator partnerships, launch videos, AI workflow building, like if we have all these and there’s like a team of like five or six agents for each one or something. We’re going to need an orchestration layer that not only manages itself, but also gives us really easy visibility into what’s going on. So I don’t know what implementing that looks like, but I do think we should consider it as we build all these out. Ben Ornstein: I agree. I think just having the visual dashboard of, OK, this is going to be LOR’s specific agents. Here’s what they’re doing. And I don’t know what the actual UI of that to look like, but I can picture what it would look like and how that would be helpful. I don’t know how like configuring all of it would work and all that. Michael Houck: Yeah, I mean, here, I’ll just, I’ll pop open the paperclip page again, because I think it shows a little bit of the UI. But like, yeah, like something like this, like, where’s the org chart? That’s what I really wanted to show. Yeah, like it has this in there, right? So like, you just super clear like, okay, you know, we have Haukbot, which is the CEO, right? And then we have, like, some, like, megaphone orchestrator, right? And that manages Casper, Eddie, Trent, et cetera. Maybe not Trent, because Trent sort of goes across different niches in theory, but you know what I mean? So, like, we’d have these teams set up, and each one would have their own roles, and they’d understand how they can relate to it. It’d just be easy also for us to visualize like a team of like 50 agents versus having to just know who to tag in Slack. Ben Ornstein: Yeah, exactly. No, I totally agree. It’s already getting crazy. It’s like, okay, I have to manually be like, okay, we gotta add Casper, we gotta add Eddie. We also now gotta add Trent everywhere. I was so pumped about the Trent output for proactive engagements this morning. I was like, yes, it’s working. Michael Houck: Yeah, dude, you really done a great job, like objectively. And I’m excited to see how we can take it from here. Because I think that like, you like one shot at that this weekend, like you wrote up the long doc, we talked it through, right, we identified the holes to poke. And then you just made it and you one shot at it. And like, it’s gonna work. And it’s fine. Ben Ornstein: So yeah, it’s, I spent probably like 10 hours this weekend on all that. Like it was just, it was just, it was like making sure everything was working to and it made sense and docs were filled up. And then like, if things weren’t working, being like, why isn’t it working? So it wasn’t exactly a one shot, but like in terms of actually building him, once we went through the whole process of being like, okay, here’s what we gotta consider here, et cetera. Like it was somewhat straightforward just given with time spent, which was good. Michael Houck: One thing we’re gonna need to do when you get back is, you know, we have these ambitions to launch all these other agencies, these other verticals, right, that we can sell in the founding journey. We’re going to obviously need agent teams for them. But in order to achieve that in a reasonable amount of time, we need to spend some time up front on like templatizing this process that you’ve gone through to set up Casper and Eddie and Trent and how they interact with each other, right? There needs to be some sort of like architecture for if we want to start building agents for the GEO agency. Okay, great. How can we do that as a whole in like a one week to two week project? Right? Ben Ornstein: So part, I thought about that in part, I have something in the workspace, that’s like a template for a new agent build. So it’s it’s part of that. I understand it’s not the whole thing. You’re right, that will need to be something we need to like, build out. Michael Houck: Yeah, like what I would love to do, what be a competitive advantage for us too, is if we can identify a need, like this context farming thing. No one is talking about this, but it is a need that people would pay for, right? So we identify a need early. And then I would love if we’re in a place where we have these agents, A, who we can use to like set things up. Like we have like a landing page copywriting agent. We have like a landing page engineering agent, right? And we have like an ops agent that handles like set up and stuff like that. And we can literally just go to it and be like, Hey, like, let’s launch this. And then within a week, we have a team of agents set up that you’ve sort of guided and architected and prompted, like what to what to do for and I’d give it my feedback and things like that. And then we can launch it in the newsletter the next week. Ben Ornstein: And we can see if it gets any feedback, any any interest, you know, it would be insane if this becomes like my whole job, like architecting, building the agents and then managing them. Even when they’re like, not working properly, I’m having so much fun. Like it is, this has been so, like I’ve said this like 10 times now, I really enjoy this work. Like when I get taken off and have to focus on like a studio client, I just want to get back to building these guys. Cause I don’t care how much time is spent. This has been, I mean, and this is like business changing. Like if we build these correctly. And I, and I know that, like, I know if these are built correctly and we can sell these things, Holy shit, this is, I mean, the margins become insane. Michael Houck: The margins become insane, the amount of revenue streams become insane. It lets us build something really big where, you know, we can hire more people if we need to, but also we can just, you know, I would obviously cut you in on sort of the profits that we get from that, because you’re so instrumental in building this and maintaining it and all that. We can figure out exactly what that looks like as we get these other things spun up. But I just want you to know, like, I’m obviously going to be prioritizing that. But it absolutely is like it creates an incredible opportunity for us as a business. And so yeah, I’m so bullish on on it. And so I think what I need to do this week is outline like what those funnels are for these other ones, and then we prioritize them. And then on your side, when you’re back, we just need to start executing and like we need to find where the line is between you and Raphael and like, maybe Maybe Mahul, probably not Mahul, but probably just Rafael. Like, you know, do we draw the distinction between like architecture prompting guidance and actual coding execution? Or do we do it between like, oh, you know, like you’re building some of them, he’s building some of them. You know, like, I don’t know what the right cut there is, but I think we’ll probably figure it out as we go, unless you have thoughts. Ben Ornstein: Well, so from your, cause I’m slow Like I’m interacting with an LLM and I’m being like, here’s what I want to build. Here’s generally what our infrastructure is. Here’s what I want to do. How would we do this? And like, I’m talking back and forth with an LLM, obviously also writing stuff myself within VS Code. So like, I’m slow. I’m slow doing it because I’m trying to consider everything. And it’s not… Michael Houck: It’s smart to be slow. Ben Ornstein: Right. And so I’m then I’m also having clarify like if I were to build this where in the docs would I actually like edit this and why like I’m truly really trying to architect every single piece of it uh without of course giving like claude like our build I’m finding the balance between those two because there is a balance and eventually like I don’t want them to know anything we’re building so but but at the same time I can’t do it without doing that Yeah, a hundred percent. Michael Houck: Um, I think that’s, I think that’s the right balance to strike, but, um, yeah, I, I just, I want to give you more firepower. Like I want to be like, you know, use Raphael if we need to, you know, bring on some other contractor to do something like the more that we can save increased margins, save money on the things we already spend money on too. Um, it’s, uh, it gives us more opportunities to do that sooner. But the biggest opportunity we have is just setting these things up so that we can funnel leads to the right place and grow faster, right? Like, even if we don’t cut any expenses and use this stuff in like that kind of way, even if we just use it for growth, I think that the results by the end of the year, like the second half of this year are gonna be insane. Ben Ornstein: From your perspective, how has Raphael been in terms of like when he hasn’t been blocked, building out the FJ bot stuff? Michael Houck: I don’t know, because I haven’t gotten to use it yet, right? But he’s quick. Yeah, yeah. Go ahead. Ben Ornstein: What I was going to say is he built that thing a lot faster than I would have, which is not surprising. Michael Houck: Yeah, but that might not be a good thing, right? If you go in there and take a look through it, and you’re like, this doesn’t use the best, because he doesn’t know, you know OpenClaw better than he does. He knows engineering principles and foundations a lot better than you do, but you know OpenClaw a lot better than he does. And so it’s like, if he’s doing it quickly, that can be good, but we need to sort of, at least in the early stages, like he’s a quick learner, he’s got that kind of like engineering mind, but we need to make sure that he’s doing it in a way that doesn’t violate everything that you’ve learned about how these systems actually work over the last month or two. Ben Ornstein: So in one scenario, I see like, well, one thing’s obvious to me. If Studio is sold and I don’t have to worry about these damn clients every second, and my whole mind is on these things, that’s a massive difference. And I don’t want to say like, okay, well, just take me off. We can’t do that, right? Michael Houck: No, we can’t. The one, the one thing about that is like, I’ve been thinking about that. I sent them some stuff over the weekend now that I think should get us to move things forward a little bit faster. So if we keep getting good responses to that, it might be worth literally just hiring someone, you training them up when you’re back, how to do the studio stuff so that they don’t, I mean, we’re automated product engagement right now, which is a big, I think part of it, but just some like account management layer person. That we can just include in the deal so that they don’t even want you involved. Because the stuff is just so high leverage that you’re doing for us that I’m like, I don’t even want that if we can avoid it. Ben Ornstein: Things are getting a lot better in terms of my involvement. It’s more of like just like the problem clients and stuff like that. So Rena, for the most part, is now handling scheduling, which is great. UNKNOWN_SPEAKER: That is great. Ben Ornstein: I was scheduling, so I trained her on that. She’s pretty good at it. She’s asking me clarifying questions. Great. That’s been really good. Now with if I can create all the proactive engagement stuff and build that stuff in for clients, like I did, Ben, Enos becomes, I mean, her time is spent like 99% on these things. If she is given the post, and all she has to do is literally log it into a spreadsheet in terms of reporting, and then actually create doing the engagements for the ones that she has access to. She is her time was going to be spent, she’s going to have so much time to do anything, but obviously not going to be the account management layer. Michael Houck: Here’s an idea. Do you think it would be interesting for, because I agree, there’s like two big pieces we’ve got out of your plate. The last one though is when these clients have like account management questions. I get looped into that, you get looped into that. Obviously we want to minimize that for the deal. So what do you think about the idea of having Blake Stewart get involved with that? Because he’s customer facing, he’s partnerships. He’s told me that he’s down to do account management stuff. He’s interested in getting interest in the other side of the business, not with Megaphone specifically, as I said, but he said just in general, as we sort of do more, like he’s talking about the AI workflow agency and things like that. But even before that, he was like, if you need me to do Megaphone account management, like I will. And he can do sales calls too. So part of my thinking has been like, maybe we just loop him and train him up on that. And then obviously we keep him after the deal too, but part of his time can go to service that business after it’s closed. Ben Ornstein: Yeah. I mean, I would love that. The studio stuff’s incredible. What you’ve built, what you’ve been able to build with this thing is just insane. Just on a whim, basically, just because clients wanted it. Being able to not have to worry about it every second of every day is nice. Would be nice. Michael Houck: It would be good if we have the op stuff totally automated between Rena and Venus, like you’re doing. And then when questions come up, you can stay focused on your stuff across the whole business. And we can have Blake sort of be in there to be just the friendly face, basically. Ben Ornstein: It would be great because like the biggest thing right now has been like Draper, like Michael is constantly emailing me stuff. Like, hey, can we have Jacob create a post for this? Hey, we need to engage from all accounts on this post. So I have to think about every client’s voice all in and I get it, like they’re paying us a crap ton of money. And it’s a great client. But when things like that happen, like hours of my life, are just gone. And I know you know what I mean. Michael Houck: It would obviously increase our expenses a little bit, because I’d have to compensate Blake for it. But he’s expressed interest in it. And it’s just so much lower leverage stuff than what you’re doing. I mean, obviously, we can’t do it right now. But it could be interesting while you’re out, actually. It’s an interesting timing. It could be interesting for me to introduce him to folks, and then when you’re back, we sort of see how that went. I don’t know. Ben Ornstein: The biggest issue is the time spent training him, because when I did this for Andrew, it was amazing when he was finally like… Michael Houck: Right, but the job’s very different now than it was a year ago, because… That’s true. That’s awesome. You’re not gonna be handing off engagements at all. He just needs to be a friendly face who can… Who can handle questions well and hop on calls and stuff like that, right? Ben Ornstein: Yeah, I mean, that’s a very good point. It is very different now. And like him, he could be the one interacting with Jacob to be like, hey, we got to make sure this is in this client’s void. Yeah, it is. You’re right. I didn’t even think about that. It’s an extremely different job now. Michael Houck: I hadn’t thought about it either. Just hitting me as we’re talking through it. Because it’s just like, if we have these ambitions for the business, that’s priority. And I don’t want once your time to be sucked up. Even if they want like 10 hours a week or something like that, like even for the first couple of months, like I don’t want that. And so I’m trying to solve that before it becomes a problem. And rather than hire someone and go through a process and try to train them up on everything Rivalry does, it might be better to just use an asset we already have, who we like and who we know will do a good job. Ben Ornstein: And who already has context. Yeah, no, that makes a lot of sense. I think that could work. I think it could. And we could be like, hey, this is introduced to the clients, like, hey, this is going to be like the guy instead of Ben, Ben’s gonna be around, but not the guy. Nice. Michael Houck: I have a call with him Wednesday morning, right before you and I chat again. So I’ll float it to him. And I’ll see what he thinks. And I’ll be I’ll be candid with him be like, hey, like, he already knows is being acquired. So like, as that goes through, like they’ll probably ask for some of your time after that, but we still want you, like, this isn’t like we’re like pawning you off, et cetera, et cetera. I think that’ll be fine. And we’ll see if he’s interested. Cause it could be, it could be an easy solution to an annoying problem. Ben Ornstein: I know I completely agree. And you’re right. Because the role is so different now. Yeah. It’s really just like being there to pay attention and like keep a pulse on things because I am doing that for every single client. It’s and I’m, you know, that’s my priority, since they’re the biggest moneymaker. Obviously, the agents are huge, too. But it’s it just takes time. Like, I was happy to be able to work on it over the weekend, because I didn’t have to. Well, actually, it’s funny enough, I still did have to worry about clients. They were messaging me about stuff. But I mean, it’s just it’s what happens. You know, it’s I’m not complaining. Michael Houck: It’s just it’s just no, no, I get it. I get it. Yeah, look, let’s let’s give this a shot. Let’s see how it goes. I’ll talk to Blake. I’ll see if he’s interested. Because the business is evolving and it’s changing and your role with this agent stuff is expanding and the importance of it is growing now that we’ve sort of proven we can do it. I mean, we just heard Jacob talk about it, right? Like he likes it. We can, in theory, change the model with the new version of this that we build because we have these agents. So it’s just getting more and more important. I don’t want you to get sucked into a whole bunch of other stuff. So, as long as that makes sense to you, I’ll float it to him. Ben Ornstein: It does. And I can, of course, still be a backstop like I was for Andrew. Like, if there are things that, like, I have… Yeah, I don’t need to explain it anymore. Yeah, that sounds great to me, honestly. Michael Houck: I love it. That’s, like, a good brainstorm idea there. Ben Ornstein: And while I’m out, he can be, even with less context, he can be someone that can, like, help Reena and Veena’s hand out. Things. Michael Houck: Yeah. And I’ll still be there to solve like the big problems. But I think just like, introducing him to clients, like, it’s actually crazy how good this this week’s timing works out for this, right? Because I can really just go to the class gonna be like, Hey, Ben’s out for this week. This is Blake. You know, he helps out with stuff across the business. And, you know, he’ll be your guy this week, right? Like, it’s so easy to do that. Ben Ornstein: That’s a nice transition to having him actually do it. Yeah. No, that’s actually, it’s actually really good. So thank God I’m getting married. Yeah. Yeah. No, that’s, that’s cool. Actually. That’s really good. Michael Houck: Um, okay. I’ll talk, I’ll talk to him. Ben Ornstein: Uh, I guess two more things that I wanted to, uh, so back to on the Raphael stuff, um, in my head, I still think it would be cool if Raphael was the one actually building and when he has like clarifying questions on context, um, See, I have two sides of it. There’s one, it’s like, okay, he builds it, but he missed, completely missed a part of context that I would have known that he definitely wouldn’t have. And that completely screws the entire buildup, maybe. And, or I’m just still doing it. Um, I don’t exactly know how to root block Raphael and cause even if I’m building some agents and he’s building others, like, ideally that makes sense. But if he’s still screwing up context, then what’s the point, you know? Michael Houck: I think we have to do what we did for Trent, which is you like you and I align on like, here’s the agent team that we need. We prioritize which ones we want to build first. We get some idea of what their capabilities might be. You do the research into planning out the architecture of the agent. You write up a doc, I go in, we go back and forth on the doc. Once the doc is what we think it should be, then you can hand it off to Rafael because he can do the execution. He just doesn’t know. We should be over-explanatory in these docs because this is the case all the time with or even with Rina in some cases, right? Like, they just don’t know some things about, you know, how these kinds of businesses work that we just need to give them context on. So we should assume that they know less context than you or I would for each other. Ben Ornstein: Yeah, that’s interesting. If we gave them that Trent doc, like another version of that, but extremely in-depth, way more in-depth that has a lot of the context for the new build, it won’t completely solve things, but that will definitely make things easier. Michael Houck: I agree. I think that that’s very, very, very straightforward to do. Obviously, it’s like a, but it’s not even an extra step because we’re doing it already between each other, right? Ben Ornstein: Exactly. We still have to do it. We still have to do it anyway. And I like that we did it because it made the trend build easier because we talked about trend spot and we talked about outliers and formats, I was like, okay, I generally know what to do, but in terms of actually building, I was pretty lost before that doc. Michael Houck: Okay, so maybe we test this out with the lore agents, and I can send over some initial thoughts on like, here are the agents I think we’ll need for this, because Emma was saying they’re like building out some like AI dashboard thing, like they’re using these tools, they’re using LLMs for like like stuff that I personally don’t think is that like high leverage, right? It’s like, it like makes it look polished and nice, but like it doesn’t actually reduce costs or allow more growth or anything like that. Ben Ornstein: Well, a lot of people do. Michael Houck: I know. So I think, I think if we find those like high fulcrum points, we can, we can do that. And also I was talking to Tatum the other day and we have a full list of the agents that we want for like a TikTok IG reels. So maybe we start with that one, actually, because that’ll just be better defined. But we could we could just do that we could just pick one or two to start with. We could do the research, we could write up the docs, we could edit them, then send them to Raphael, and see what he comes back with, basically. Ben Ornstein: Yeah. And so I guess he would do the same thing, right? Like he’d be SSH into the Mac Mini, and he would just be able to just make changes. Michael Houck: Yeah, Yeah, maybe there’s some like, maybe we set up like a GitHub structure where it’s like, you need to review his pull requests for when he makes changes. Ben Ornstein: Yeah, I was thinking that, because I would love, because if he changes core architecture without, because that could be a problem. Yeah, we’ll have to think about how to do that. Michael Houck: GitHub has like standard functionality. Like if we host the code based on GitHub, I don’t know how it all works, but he how it works. And so you can explain to him what we want to do. And he’d be like, okay, like, I can set that up on GitHub, where like, you’re the reviewer before anything goes through. Ben Ornstein: Yeah, that makes sense. Okay. That sounds good. On the on the acquisition side, what’s the what’s the latest there? Michael Houck: Yes. So I sent over what I think is like the final packet of docs. And they wanted like all of the agreements we have with every client and vendors and contractor agreements, like everything and the whole bunch of context on it. They were like, they were like what they had a bunch of like questions to like, like about like our growth channels and like, Hey, most of our stuff comes through word of mouth for studio right now, right? Which they said was fine, because we have viral examples, and they can run add traffic to it. And they’re excited about that. So I said to write out a bunch of stuff, but now they have all that. And so now I’m like in some portal that they have set up. And I’m expecting I shot them an email last night, too. I’m expecting to up on a call with them tomorrow or Wednesday, we’ll see. And they’ll want to go through the P&L in more detail. And they probably will also request some more information. But still moving forward just slower than I would like, Obviously, I’ve communicated to them now that we’re about six and a half, I guess, weeks out from me going on vacation. And so I was like, we need to wrap it up before then. So they know, they understand there’s some urgency. And we’ll see what they do with that information. I also told them that April was our best month ever, which it was, because we had the Draper payments go through. So they know those are annual payments, but still, it was nice to be able to say that alongside this. So I’m just trying to give them a kick in the ass, move a little faster. Yeah, yeah, yeah. Ben Ornstein: Plus, technically, for the agreement in terms of the numbers that we hit, we officially passed those. Michael Houck: Well, May is still part of it. What we need is we need to hit the same profitability in May, basically that we hit in March, March and February and stuff. So we should be able to do that, but it would be definitely be nice if we sign a couple more clients in the second half of the month. So we’re gonna work on that. And we need Dom’s payment to go through too. Ben Ornstein: Yeah, we definitely need that payment to go through. Michael Houck: But yeah, we just need to hit that number this month, which we should be able to. So I’m gonna hit the pavement a little bit later this week on some sales stuff for while you’re gone, try to get a couple people to pay before end of month. Ben Ornstein: Awesome. Okay. Well, I’m gonna, I guess, hop on the call with Jacob and James. James’s voice is crazy. It’s hard to match that. I see why Jacob’s having issues. It’s just, it’s a lot of ridiculous, but he’s a weird guy. Michael Houck: And honestly, like, it’s always the people who don’t have any history of writing who are like, so particular about how they come off. And it’s like, dude, just do it works. You know? UNKNOWN_SPEAKER: Yeah, exactly. Michael Houck: Yeah, I’ll let you go to that. And then I’ll see you in a half hour for the newsletter sync with good word. Oh, they tried. They tried to do it without Caroline. And so I Mrs. Laura and I was like, Hey, I think Caroline should be here. So y’all get on the same page. Ben Ornstein: We need to get on the same page, like, oh my God. So yes, I’m glad you did that. Michael Houck: Subtext being like, what are you doing? Yeah. Michael Houck: All right. Ben Ornstein: So I’ll hop on this call and then that, and then I’ll do reporting and then I’ll do house picks eventually. I’ll get it done. Don’t worry. Michael Houck: No, all good, man. No, no worries. I’ll see you later.