The Exchanges

Q Yeah. I'm like, have you used any other tools? Like why you, why suggest Google AI studio?

A I use AI, so AI studio is my favorite for like a one shot prototype. So if you just want to see something, you're not working in an existing code base, you're just like, what would this look like? Google AI studio is really good because it'll do it end to end. When I'm actually working day to day in an existing code base, um, like in my current startup, we typically focus on using cloud code just because it's really good at working within a code base. Um, and then sometimes if I do want the capabilities of Gemini, But I still want to work in, um, an existing code base. I'll use Gemini's model and cursor. And then I found that Gemini is really good at UI and shader work and like more complex animations.

AI assessment note: “AI studio is my favorite for like a one shot prototype.”

Q Like I feel like a lot of people from this, you know, listen to the show, come steal an idea. They're doing a hundred K a month. Um, they're feeling like they're on top of the world. Um, and then churn just gets worse and worse. How do you think about churn, and how do you, like, is there a scientific way to combat churn?

A So I think when you, it's kind of like when you're riding a motorcycle, what is the first thing they teach you is like, don't look at the wall. And I think that oftentimes the, the big four letter word and startup ecosystem is churn. And I think the problem is when you think about churn as your main focus, it ends up thinking about the opposite, which is actually what you really should be thinking about, which is loyalty. So loyalty is the way I measure it is net revenue retention, which is not just how do we avoid churn, but rather how do we get more loyalty from each existing cohort? Every three percent increase in net revenue retention doubles the company's valuation.

AI assessment note: “measure it is net revenue retention, which is not just how do we avoid churn”

Q Okay, so what is your, your vision here? So you're gonna, let's, you know, to create a business around this, like you'd actually, like how do you actually take this from, okay, I'm, I'm buying cheap, undervalued pizza ovens to like, I, I, this is an automated business that's generating income.

A So what I would do is I would have the operator go do research on eBay, uh, pricing research and give me an average price for the same make model oven. And then let's say it's 300 dollars. I would have it go make offers on my behalf for a 150 for every Gosney that's been listed for 30 days or more, the longer it's been listed, the more likely they are to take a lowball offer. As soon as they say yes for a 150, um, I would go lit. I would take the pictures, list the same thing on eBay. Um, for 250, um, at a buy now price, once it sells, I would collect payment, I would send the payment to Facebook, to Facebook Marketplace, and I would provide the ultimate end buyer's, uh, address to the Facebook Marketplace seller, and they would just ship it directly to them.

AI assessment note: “I would have the operator go do research on eBay, uh, pricing research”

Q So that's interesting, because it's like, I, I would want to work on apps. Like, when I'm thinking about consumer AI apps, for example, like, the ones I would prioritize doing would be ones that Have high shareability, and this, to me, has low shareability, right?

A Yeah, definitely can have low shareability. It might get to the point where someone finds it very useful, their friend tells them, hey, I have this weird bump on me, and then you say, hey, I know this app, let's just take a picture of it and see what it is. That's something where I could see it. But where I really, and I could see it being also shareable through word of mouth among parents. So not sharing the actual result, but sharing the idea. Parents, concerned parents are always talking about their kids and whatever's happening in their life. So I could see parents sharing to their parent friends like, hey, your kid is feeling sick. Just get this app. It's what's helping me. And it's spreading that way. But I think that the primary lever to drive downloads to something like this would have to be paid ads or influencer marketing, where we could get into it more later, but it would play on the emotions of a little bit of fear mongering on the idea that they might have some kind of disease or problem.

AI assessment note: “Yeah, definitely can have low shareability. It might get to the point where”

Q And so I was hoping that you can come on and you can share some of the most interesting trends, topics, and ideas, and you're a listener of the pod, so you know what this audience wants. Uh, Brian, what do you got for us?

A Yeah, the first idea right off the bat is one that I would pay for. It's a take my money situation, is an AI content updater. So there are a 1,000,001 tools for creating content with AI. There's, you know, content writing tools. There's, uh, tools that do outlines. There are tools that rewrite your content. There are tools that'll take your long form video and chop it up into shorts. But what real publishers need is help updating their existing content library. So for example, with exploding topics, we have maybe like 500 blog posts, which is nothing compared to most publishers. And even that we have trouble keeping up to date. We have about, uh, you know, one and a half full-time people constantly updating this content and they're going in, they're saying, okay, this stats a little bit old, or this isn't really as relevant anymore, or there's a link that's broken, or there's a source that's more recent we can use and they're doing it manually. So if there was a tool that could go through your existing content library, find areas that need to be updated and actually go into your CMS and update it, it would be a game changer.

AI assessment note: “the first idea right off the bat is one that I would pay for”

Q app around that and they have a whole community and a brand. So he was able to take that and turn that into a brand for a specific niche within meditation. That is this like skeptical, uh, group. So that's one that comes to mind as built completely from scratch in a growing space. Um, that's one. How about you? Do you have any Greg that you can think of?

A The one that comes top of mind is a friend of mine and friend of the pod. Her name's Natalie Ellis. She started, it started off as like basically a meme page called girl boss. I think it was called. Um, and she was sharing memes and quotes and stories around female entrepreneurship, women led businesses. Uh, And it grew to 6.4 or something million followers on Instagram. Big, you know, and once she had that, she was like, okay, what are the things that I could sell? Templates, uh, a membership, uh, you know, events. Uh, now she has one of the top podcasts, one of the top business podcasts. So, um, she went from a meme page to a millionaire basically quite quickly. And, uh, it started with the brand.

AI assessment note: “The one that comes top of mind is a friend of mine... Her name's Natalie Ellis.”

Q rank high in Google. That's why I like SEO, and that's why I use BoringMarketing.com, and that's why I invested in it. They're so confident in their approach that they offer a 30 day sprint with a hundred percent money back guarantee. Who does that nowadays? So check it out. Highly recommend BoringMarketing.com. Alright, next idea, uh, Number four, let's, let's, I like this idea. Tell me more about it.

A Okay, so Shady Rays is a hundred and thirty million dollar, um, sunglasses brand, and their entire business was built on their lifetime warranty. And so forget like Patagonian, what all the other brands do, because that's a different expensive product. Shady Rays said, we know that you're going to break or lose your sunglasses. Don't worry about it. You know, we will replace them, no questions asked. So it's not like, you know, you have to return them if they're broken. It's If you lose them, we will replace them. And then they work like a, a restocking fee into it to where they, they break even on all the replacements. They work the numbers into it where they're like, all right, you know, most people replace their glasses .7 times over their lifetime. So, and then they, the, the lifetime warranty allows them to charge a lot more than a competitor and they crush it. So I just love the idea of copy pasting this exact value proposition in any other industries. Such as sandals, like flip flops, lifetime sandals. Um, same thing, right? You can go to AliExpress and buy nice looking flip flops for three to four dollars. Shipping is 50 cents to a dollar. Sell them for 50 to 70 dollars. You have amazing growth margins. You charge five dollars, um, every time someone needs to replace them. The five dollars covers the three dollars in shipping and the two dollar sandals, and you're great…

AI assessment note: “I just love the idea of copy pasting this exact value proposition in any other industries.”

Q I don't want to waste a second. What, what's, what's your first idea for us?

A Okay. Idea one. K-pop style factories for Particular talent. When you're building a startup, there are certain unicorn hires you just have to make, but every single startup has to make the same unicorn hires. Like here's an example. When a startup hires their first growth marketer, that's a specialized hire. It is a hard hire to find. Like our first marketer, we spent eight months searching, like really searching, but that's a set of skills and you can teach it. And so you could imagine a world in which there's a K-pop style factory where they are churning out the first growth marketer for startups. And people will pay a lot, and they'll, they'll, they'll pay to not have to spend too much searching and to have the person start the next day when, honestly, they needed someone yesterday and didn't start hiring soon enough.

AI assessment note: “Idea one. K-pop style factories for Particular talent.”

Q And with your, with the lab and specifically, what are the price points and how many people are taking it? And how much revenue does that bring in?

A I've taken some interesting design choices with the lab. Uh, we have 200 members because that is the cap that I put on it. When we got to be about 50 members, I said, I'm actually gonna keep this at 200 people. Um, so we, we hit that 200 member cap at, uh, in February of this year. Um, there were two tiers of membership. There was a standard tier, which was community access and access to all the courses. Then there was a VIP tier, which was The standard tier plus a quarterly one-on-one call with me. Uh, the standard tier is priced at 1999 and the VIP tier is priced at 29 99. So that part of the lab.

AI assessment note: “we have 200 members... standard tier is priced at 1999 and the VIP tier”

Q And where's video in this whole ecosystem? Like, You know, everyone's saying video, video, video, short form is everything. Like, how do you think about video in your whole Morning Brew ecosystem?

A For us, short form is important, but it's hard to monetize. And so you have to do short form in a really inexpensive way. And so either you need to take long form. And make it in short form, right? So take a podcast, clip this up, and put it all over TikTok. There's a bunch of people out there who do that quite successfully. I know a few who we've worked with. That's option one. Or option two, make it at basically zero cost, right? And so make it with just a great personality and a camera. But it's very challenging to monetize short form via ads at least, because how do you integrate an ad partner into a 32nd video?

AI assessment note: “For us, short form is important, but it's hard to monetize.”

Q So did you, did you know Tiago or you basically just took his course, got to know him and then he became your business partner?

A Mostly the second. I just really admired Tiago. I still do. And I was like, wow, this guy understands something about information capture, something about sort of the creative process, something about online education that I would really like to learn from. And the man and his system just radically transformed my life. And so I reached out to him. I remember we had lunch in San Francisco. And I called him on a rainy day in New York City, December seventh, 2000 and, and 18. And he, in a five minute conversation said, sure, um, I'm happy to do it. And at the time I had very little money. And right around then I got a 20,000 dollar grant from Tyler Cowen's Merchant Ventures program, took that 20,000 dollars, hired a Hollywood film crew, flew them down to Mexico City. We made these Seven video modules, and in 10 days, came up with all the concepts that would eventually form the bedrock of Rite of Passage, and here we are now.

AI assessment note: “Mostly the second. I just really admired Tiago.”

Q And over here, I see you have site builder stack. What do you mean by that?

A Yeah. So I, I just have an agent on my other account that has this simple stack of cloud flare. Um, the CLI tools, uh, pork bun is where I buy domains so it can log into that on its computer and buy a domain. Um, and then I use this thing called Astro for S for it's a framework for creating static sites and it's connected to my GitHub. And so basically like I needed I needed a link for people to book meetings with me. Uh, and I didn't want to pay for Calendly. So I said, I got it set up on this stack. And I said, go embed my three different Google calendars on this. And I was literally sitting on the phone on the beach and it would send me, okay, here's a mock-up of this. And I would tell it, okay, make another branch and do a mock-up in this style. And it will come back and through Cloudflare, it would have separate Developer links that I could review the different styles in. And then when I said, okay, I like that one, they connected it to the domain that I owned and it went live all from the beach. And I set that up in Two or three hours that, that, that stack. Um, and I'll probably use it for a long time just for simple. I need something to exist on the internet. It doesn't need to be perfect or pretty.

AI assessment note: “I just have an agent on my other account that has this simple stack”

Q I mean, even, even in a world where there's the maintainer, if something breaks and you're an enterprise, you want someone to call. You want to go into someone's office, right?

A Yes. You also want someone to blame. You want someone to blame. That's an important piece. I think a lot of people are forgetting that, like, The, the question of, is AI gonna replace this, this, this, whether it's a task, a job, a company, a product, something. Um, often I am asked, the first question I'm asking myself is, who's liable now? Who would be liable in that other world? And do I think that that trade-off is worth it right now? Like, I work with Fortune 500 CEOs every single day. They, no way. No way. They want to be able to call because they want someone to unblock. They want someone to secure. The other thing is that, um, let's say that, um, let's just say it's a Salesforce example and that you could build a shitty CRM or a simple CRM or something that's just running on your own. Um, but Salesforce has relationships with all the AI labs. They are, you know, getting into early testing. And so by the time a new model comes out, You are facing it as a day one person. They're facing it as a day 30 maybe. And so you're also going to be on a very big lag. Um, and so as you're thinking about that cost trade-off, I think in addition to all the things that we just talked about with enterprise-grade security and maintaining whatever, you just also don't want to experience that lag. Like we're moving to a world where being fast to the punch and getting a 30 day, 60 day, hundr…

AI assessment note: “Yes. You also want someone to blame. That's an important piece.”

Q Um, not affiliated, but you can just see like what's charting and what are people downloading? And it's like, okay, in a world where super intelligence is now on tap, How can I make an AI native version of this? Um, or undercut, you know, from a price perspective or just drive more value. Like there's ways, there's now like opportunity to, to, to, to enter some of these markets.

A I, I completely agree with you, and I think Agent First software is absolutely one. Um, two things that I actually think are really interesting, or maybe three by the time I get to it, but interesting research avenues to learn more opportunities like the one you just mentioned. So one, YC posts videos on Instagram for what type of applications they're looking for, and Agent First software is one of them. So Listening to what YC is asking for, assume that they are already thinking, 18 months out. Um, so that's definitely one arbitrage research opportunity. The second is Matt Van Horn's last 30 days research skill, which is just amazing. I've, like, inner, um, I've integrated that with my, like, Claude Wiki. Love it.

AI assessment note: “interesting research avenues to learn more opportunities like the one you just mentioned”

Q Um, and Cody, can you explain like what, What is a data warehouse? Like what are these things?

A Yeah, absolutely. So a data warehouse is a way to unify all of your data sources, um, into a single location so that your agent has the ability to understand all of the data in context with all, like all channels in context or all data sources in context with all the other data sources. So I'll give you specifics. That was a very like, you know, broad way to explain it. So say, for example, I'm running Facebook ads. How do I actually run effective Facebook ads? I need to do probably four things. I need to have my Facebook ads data. I need to have my Google analytics data. I need to have probably my post hog data, maybe my, my CRM data. So like my HubSpot data. And then I also need to probably have like my Stripe data to understand the entire stack of like, I ran ads and did customers come out the other side. Right. And then how do I actually tie The literal specific ad that is driving revenue for the company, right? This is how you do that. It's by having this data warehouse that has all of the data in a single place. Now to get all of that data into a single place, you have to build a data pipeline. And so Airbyte is a data pipeline solution. It's self-hosted, um, and it enables you basically, they have all of these prebuilt connectors that enable you to pipe all of your data Into a data warehouse. So when you think about like, you know, visually how this is happening, and we …

AI assessment note: “a data warehouse is a way to unify all of your data sources”

Q tasks that you as the human being are going to be doing or to the sets of tasks and what are the sets of tasks That the agents are going to be doing that you're okay with them doing and keeping you in the loop. So my question for you around automations is beyond, beyond QA, what are some other automations that, you know, a, a founder could be implementing?

A So, um, let me pull up my list for you. So the kind of automations that I have are, uh, clearly, uh, uh, QA bug fix fine. Um, the other is what I call my production watchdog. Um, so we've got a lot of real law firms using Untangle. We've got a lot of real clients, you know, going through discovery and we've got, with just a huge amount of activity. And, and honestly, I can't keep track of it, uh, you know, mentally by just looking through logs. And so what I do is every day at nine AM, I have what's called a production watchdog automation. Which what it does is it goes through all of the events in the database that happened for our paid customers, and it summarizes that in a JSON file. And then that JSON file is in our admin, and then I, you know, know every morning I go in and I click on the production watchdog and I read through the summary of what the customers did. And it's so valuable. Um, so it's sort of like you can imagine, you know, your chief of staff showing up and saying, you know, all right, Greg, here's the important stuff that happened yesterday for our customers.

AI assessment note: “The other is what I call my production watchdog.”

Q also, like, it's easier to sell monthly if you've built trust and confidence with them. So sometimes you have to do a little Like show them, you have to show them, Hey, I could build a skill here. Oh, it's starting to work. Okay. You want more of these? Well, you know, for a thousand dollars a month, I can give you two of these. You know what I mean?

A Yup. Well, and that's the beauty of the assessment too, right? Like these, we're not pitching these cold. Like these are people that are already clients. So it's a lot easier to pitch a, you know, a 5000 dollar clawed skills pack to somebody who just paid you a thousand dollars. Then it is to pitch that to somebody that is like cold traffic off a Facebook ad, for example. Right. And another just side note, I forgot to mention this, but one thing that I love to do now, I usually don't lead with this because I don't want to do this if I don't have to, but a really good way to get an upsell over the finish line and get a prospect off the fence is if you already did an assessment for them for 999 dollars, you could say, hey, How about I just credit the 999 dollars you already paid me towards the implementation, right? So instead of five K, it's four K. So basically the assessment's free. Like that works so well. And what I really like to do is mark up the upsell an extra thousand dollars and credit the assessment towards it regardless. So that way I'm still making the same amount of money, but psychologically they feel way better about it because they're like, oh, I'm getting a thousand dollars off. Actually you're making a dollar more. Yeah, technically it's profitable.

AI assessment note: “Yup. Well, and that's the beauty of the assessment too, right?”

Q on. People forget that. People forget that in services businesses that there's always a more luxury, luxury, luxury experience that, A certain ICP is going to want, and by focusing on this group of people, you're kind of missing out on, on those. That being said, you have to take it one step at a time, right? You can't do all at the same time and expect to have everyone.

A A hundred percent. Yeah, and focus is key, and that's why, for most people, what I tell them is, you know, do the assessment, because that's relatively easy to fulfill, and then choose one of the upsell options that we mentioned, right? I, I recommend AI concierge, but you might say, oh, I want to specialize in knowledge basis, right? So I do an assessment and, or, um, knowledge systems, right? We do an assessment and then we build knowledge systems, or we do an assessment and then we build Claude skills. So again, that's the offer in a nutshell. What makes it feel premium? Like I said, that, that pre-engagement, uh, form, and then we've actually got a one pager in notion where all of their, All the whole client engagement lives. And in particular, we have an inventory of everything we've accomplished together during the engagement. So that is what's staring them in the face that says, Hey, we did this, this, this, and this, this month. Okay. Yeah. I'm going to keep paying Corey, 1800 dollars because these things are impactful. So that notion hub is always updated after every call. I email it to him after every call, we make it front and center. And that is the AI concierge engagement. And that is, that's the full stack, man. That's everything. We just gave it all away.

AI assessment note: “A hundred percent. Yeah, and focus is key, and that's why”

Q They would sort of, it was more artistic. They would kind of like just put out a product and change it, you know, as they go. But I think the, you know, Toyota slash lean manufacturing process applying that to startups was one of the reasons why we had such successful startups post, uh, you know, in What are we talking about when we're talking about loops with AI agents?

A Yeah, so I think maybe the best way to explain it is to jump into that first example I spoke about. Well, I guess let's talk about a sort of a loop, loop engineering first, and then jump into specific examples. But I would say if the loop for the Lean Startup is build, measure, learn, we have very similar steps here with an agent. We have this build step, which is like telling an AI, hey, go build me my new SaaS, for example. Then we have this verify step, Where, you know, if you're building product, the verify step might be that all tests work, or the agent has used a browser to make sure it can click through everything, or there's some other agent that's running. If you're doing this in Claude code, and you use slash goal, so that's basically running this loop, and it's got this other agent checking, has it actually finished or not? Is it working? And if it's not working, it's just going to tell the sort of the main builder agent to just Keep looping and looping and looping till it's fully working. Some other examples of this within engineering. So you always have this stop condition. You don't want the AI just to loop infinitely. There needs to be some sort of result that you converge on. Um, so at some stop condition examples, one is like the feature works in the browser. I want signup, like the goal is to have signup working, you know, you slash goal, make signup work. And…

AI assessment note: “if the loop for the Lean Startup is build, measure, learn, we have very similar steps”

Q listening to this is like, okay, great. The, you have all this amazing context because you have like a team of 55 people or whatever of some of the smartest people in AI and product. I don't have that context. So for people who, you know, don't have that context, but who want good You know, good outputs, be it product, be it whatever, how, how do they bootstrap context?

A The world is a large and lovely place, my friend. So we are not the only people who have produced beautiful work. I think for net new stuff, we're among the best in the world of thinking about AI flows, conversational UX, how to design for trust, especially with agents. Not a lot of people have done that. If you're looking back on this, this is go to Mobbin. Mobbin has an MCP now. Mobbin is a library of a bunch of beautiful apps, their flows, and all of the different permutations. Get Spotify's design system or another one that's similar, create a skill around it, plug into a mob and MCP and all of a sudden you can create this in minutes as well. So this is an only LCA stuff. Maybe the idea, okay, cool retention. You know, we had something called the daily five back in an early startup and maybe these ideas are easier to come by for us or faster to come by from us. And maybe our agents are more plugged into that, but in terms of producing something like this, People can do it just by using the right tools, and creating the right skills, and then slowly loading in the right context over time.

AI assessment note: “go to Mobbin. Mobbin has an MCP now... create a skill around it”

Q Can you walk me through the skills and tools section?

A Yeah, a lot of cool things here. This is a nice interface for all the skills that are installed on your Hermes agent. One thing people don't really realize is out of the box, your Hermes agent has, like, over a 150 skills installed, and each one of those actually adds context to your agent, and so that's increasing your cost. Now, with this nice user interface, you can come in And you can, uh, turn off any skills that aren't necessary, and you can also see the new skills being generated based on the actions you take. Not many people realize this. They know Hermes is very self-improving, but they don't realize it's actually creating tons of skills in the background while you're not looking, right? Like, I have a three JS one file game, because I build a bunch of three JS games just for the fun of it. Uh, YouTube new scripts. Real-time stock dashboard. Like, there's a lot of skills that get built here, kind of, in the background as you're doing things, and you can just turn off, turn on any ones that aren't relevant to you to save money and increase efficiency. Tool sets are groups of tools. So this is a new kind of concept being rolled out right now where, uh, you can group together many skills and tools all in one. So if you're doing something complex, it groups them all together. Uh, I got to play around more with this to be quite honest with you. Uh, but tool sets, something …

AI assessment note: “This is a nice interface for all the skills that are installed on your Hermes”

Q not even be this idea, but you might be like, you know, I'd like this, I, you know, uh, but if it was in another market or if it was an app instead of this, but my question to you is in terms of local models, is your recommendation that people go out and buy, um, like what should they be buying basically and why should they be buying it?

A My recommendation up to this, or up until recently, was the Mac Studio. The Mac Studio, from a speed and bandwidth perspective, isn't as strong as NVIDIA hardware like the Spark and other chips, but you got a lot more unified memory, which allowed you to run much cooler models and much better models, just a little bit slower. The challenge is, is, uh, Mac Studios pretty much sold out on, uh, any, basically every level. You can get it right now, but it's small amounts of memory. It's not like the big, uh, the big boys that allow you to run really good models. So now, I think my recommendation for most people, unless you can find, you know, a Mac studio secondhand or something like that, uh, the DJ X sparks a really great plug and play device. You, you just plug, you don't even need a monitor to set it up. Like you just plug it into the wall and it works. Um, it has a 128 gigabytes of unified memory, which basically means you can run any model that goes up to a 128 gigs. So when, uh, I think 27 B Works really nicely. Any of the new Nemo Tron models. So NVIDIA got into the model game recently. Now they're putting out their own open source models. They're building it with the spark in mind. So I think if you want plug and play something to run inference for you to sort of run your models, the DGX spark is a great option. Uh, I prefer the Mac studio because It's running on the devic…

AI assessment note: “my recommendation for most people... the DJ X sparks a really great plug and play device”

Q Um, and so what we can do. Like, we would put this in the eval, right?

A Yeah. You could do both. So one is, like, you could immediately go and turn this, like, or update the skill based on this feedback. You could also have it immediately just, like, turn around, like, a new draft of these tweets, right, to sound more colloquial. And then finally, to your point, I could go and create a rubric that actually says, like, okay, like, here's the five dimensions I care about, and then auto-evaluate every future output, right? Um, so, You kind of have a number of different options, like, depending on how far you want to go right now. Like, if you just want to get your job done right now, you don't want to bother with rubric, you don't have to, right? But eventually, like, you get to the point where you want to set up a scalable system for this to just constantly work and get better and better, and that's the point at which you would do a rubric, which is not that hard, actually. Like, it, you know, you can either go in through the UI and build one, or you can actually, in this chat, like, say, help me build a rubric to score great Greg-style content, which I'll queue up for after It updates the skill, um, and, and then it will go and help me create that rubric, save it, pin it to this agent or to this skill, and then automatically run every future, um, uh, time I create content.

AI assessment note: “to your point, I could go and create a rubric that actually says”

Q know, installed OpenClaw. Um, You know, I, I'm just, what I'm begging to know is like What I don't want, actually, I should restart. What I don't want to have happen is I install Hermes, and then a week later I go back to Open Claw, basically. You know what I mean? Like, I just kind of want to pick an ecosystem and be like, this is my personal agent.

A No, it makes a lot of sense. Um, and that's kind of the problem that I was having before. I kept switching and jumping between different agents until I landed on Hermes, and I've been using it for over three weeks now, which in this space is a lot of time. Um, So I'll, before we get into the installation, I'll walk you really quickly on my setup. So when you just type in Hermes in your terminal after it's installed, um, it's gonna try to open up my Xcode MCP, which is not running right now, so it'll, it'll fire up in a second. Um, cool. Yeah, and I'll zoom out a little bit. You can see here, um, when you open up Hermes for the first time, on the top you'll see the available tools. So Hermes comes built in with 40 plus Built-in tools that OpenCloud doesn't have. So you don't have to go find out which tools to install. Many of the built-in tools will cover almost all of the basic tasks that you'll need to do. If you want to fire up a browser, if you want to search the web, if you want to create recurring scheduled cron jobs or like scheduled like code, um, it has all of that built in. Even things like image generation are built in, home assistant capabilities. You don't really have to configure or go figure out what the best Tools are. Um, the other thing you'll notice about Hermes when you first install it is that it has a lot of the really most popular available skills pre-inst…

AI assessment note: “Hermes comes built in with 40 plus Built-in tools that OpenCloud doesn't have.”

Q Okay. What are, what are other must know things about Hermes Agent?

A You do still have to update it every night. It's still technically beta software. Um, so you can see I haven't updated this one since in nine days and I am 535 commits behind, which is quite a bit. Uh, so you do still have to update it every single day. It is still technically Beta software. You still should probably constantly, you know, lock it down in certain ways. Um, a really simple way to lock down Hermes agent, but still have access to it from anywhere is one to set up like telegram or WhatsApp. Um, another thing that I highly recommend for any of these tools is that you install tail scale and you configure tail scale correctly. So tail scale will allow your phone and all of your computers to be on the same kind of virtual network. Um, and then you can remotely access them using any, like, terminal SSH app. Which just lets you kind of remote in and, like, monitor it and chat with it that way as well. Um.

AI assessment note: “You do still have to update it every night. It's still technically beta software.”

Q Um, now in my Open Claw setup, I actually, it still works for me. So I'm not really sure what the deal is. Like, why is it working for me? And so I, I realize that you've recommended Not to use Anthropic as the primary model in this example, but, you know, how should people think about using a model that's been officially banned by the company that makes it?

A Yeah, that's a great question. So, and that's actually, actually exactly the reason why I put, uh, OpenAI as my number one. Um, they, because OpenAI has stated that they're okay with it. Um, so for them, it's definitely fine to use this OAuth method with Anthropic. It's kind of a gray area a bit. I think officially in their terms of service, it says you're not allowed to do that. Um, But, um, then there has been this, this statement by one of the engineers saying that it's kind of allowed. So I don't really know what the, what the answer is. And some people have been getting banned over it. So I do recommend that if you, if you're scared about your account getting banned, just create a new one, create a new account and put the 20 dollar plan on that. And then if that gets banned, it's, it's not, uh, the end of the world.

AI assessment note: “if you're scared about your account getting banned, just create a new one”

Q That this is really like a new way of working, right?

A Yeah, it is. It is. It's so new, and I, and I, even this task, it's really simple, right? Just sending an email based on a call with a proposal link and stuff, but like, even if you can just do something like seven times faster without having to go into all these tools, copy the meeting notes into the page to give it context on your meeting, it really starts to compound. Then you start to fit like a week in a day, and then seven weeks, In a week. Um, and stack that up over a year, and you're gonna be miles ahead of everyone else. Uh, and when we get into skills, you're gonna see how this continues to get even better. But, um, you can see here, it's drafted the email, it's pulled in all these insights from our call in Granola, which is like where I do my meeting notes, and then it's created the Stripe payment link. Here, ready to go. That's cool. And then now I can just go, um, send this email. And it will use my Gmail integration to, to go and send it. Uh, and then.

AI assessment note: “Yeah, it is. It is. It's so new, and I, and I, even this task”

Q Ok, so wait, so, how did you create that skill?

A Ok, so, there's two ways that I find useful to create skills, is one, you can have an idea of a skill you want to create off the bat. So, like Viral Hooks for example, I had this course on Viral Hooks, which I transcribed, put it into Claude, and Claude has this by default, it has a skill creator skill added into it. Same with all of these major agent harnesses, they'll have a skill creator skill. So it's kind of like skill-ception. You use the skill creator skill, and you say, hey, take this course on viral hooks and create a viral hook skill, and it can create it like that. That's one way. Uh, and then it will package it up nicely with that skill.md. It'll do the whole thing for you.

AI assessment note: “I had this course on Viral Hooks, which I transcribed, put it into Claude”

Q I love it. And for the beginner, are you, like, did you recommend people, you know, use OpenClaw or should they be using Cowork or Manus and some of the ones you showed?

A So, great question. Uh, I would say that OpenClaw is probably, like, one of the hardest to learn and set up of these harnesses. I would say Cowork is probably the easiest. I think Perplexity Computer, you did a video on it. Um, it's pretty simple too. Same with Manus. Um, but I would definitely learn, um, and get comfortable using, like, Claude Code or, um, one of these other ones before I started to play around with OpenClaw. And I would also, uh, have all the processes built out in Claude Code first. So, for example, that executive assistant, over the next, ah, two weeks, I might build out a bunch of skills, like the Sebastian Refer skill, um, like a daily brief, meeting prep, Et cetera, et cetera. And then once I'm happy with how it's all functioning in Claw code, then I could look to migrate that into OpenClaw where it has that more autonomous nature to it. So that's kind of how I think about using OpenClaw and those other harnesses.

AI assessment note: “I would say Cowork is probably the easiest. I think Perplexity Computer... pretty simple”

Q Sorry, I have a question. Like, when you say he's gone and found this, or something has happened, is he texting you? Is he using, you know, Telegram? Do you have mission control? Like, how, how do you, how, how are you communicating with him?

A So I, I don't really believe in the mission control stuff or multi-agent. I just have Larry as the, the one agent that I text through WhatsApp and we just message like you would an employee. So nothing fancy, uh, Larry, what are you generating today? Take a look back through the, the previous winners. This was back then is all automated, um, in the Larry scale now that he will every now and then go look at your TikTok analytics and find out what the winners were. But this was the stage of, of building it and building the system, finding out what worked. So it was still very much me saying, hey, go look at, go look at the TikTok analytics.

AI assessment note: “I just have Larry as the, the one agent that I text through WhatsApp”