The Exchanges

Q So can you, and can you just talk about Playwright and where you use it, how you use it?

A Yep. So Playwright is browser automation, and when you use it in an MCP, it allows Cloud Code to go and open up a window and do browser actions. So it can click through it. It can take screenshots of a website. Um, it can get all of the information, uh, that you need, say for design inspo or something like that. Let's say you're doing like Branding, or you're doing design, uh, you can easily access websites right here. So, in the past, before I used the Playwright MCP, my desktop was littered with screenshots, and I'd take a screenshot, I'd drag it into cloud code or whatever, but this really helps, and it feels smooth, kind of like when you're using it with voice, you can just kind of command, uh, what you want it to do, and, uh, you know, we're going out and analyzing competitors right here from a voice prompt.

AI assessment note: “Playwright is browser automation, and when you use it in an MCP”

Q So this isn't just, um, an LLM, you know, making stuff up, right? It's actually true. This is, this makes sense.

A Yeah, this is, uh, this is not just an LLM making it up because what we're doing with this MCP integration to the workflows and in it in is we're pumping all of the right contexts into Claude here, you know, so basing all of this around real data, uh, that's out there. So there's no hallucination. It's not just making things up around, you know, what they might think, uh, people are interested in. It's actually Understanding the demand. So it's finding the keywords. And then we selected a group of keywords and we're like, Hey, go out there and read all of the ranking articles, posts, and websites on the front two pages of Google and tell me exactly what's missing. So this gives you like such a targeted spear. If you're a newsletter platform to know exactly how you can go and attack the market.

AI assessment note: “this is not just an LLM making it up because what we're doing”

Q Or if you had some sort of lead magnet or CRO, conversion rate optimization skill, you could invoke that, right?

A Exactly, yeah. And to do that, you could say, hey, um, perplexity, go do deep research on lead magnets and conversion rate optimization, uh, or whatnot, uh, compile all the learnings, and then you can say, hey, Claude, create a skill based on this. All right, so we're building it now. And then after we build this lead magnet, we're going to get to sort of like the traffic generation part. So you've got a landing page, you've got a website, you've got unique positioning. It looks good. It feels good. It sounds good. You've got this mechanism to collect email addresses, and then we're going to, we're going to figure out how we nurture those leads and we're going to figure out how to drive traffic into the whole system. And again, complete marketing system. So far we're at I don't know, 54 minutes or something like that, and, ah, you know, typically this would take a lot longer or it would cost a lot of money, ah, to hire an agency to do all this for you.

AI assessment note: “Exactly, yeah. And to do that, you could say, hey, um, perplexity”

Q Do you ever use Uh, Claude on, or Claude code on the phone, iOS?

A I don't. Uh, I tried to do it one time. Uh, so I used Claude, uh, the Claude app, and you can access the same sort of, like, tools from Claude desktop in the app, and, uh, Claude code there could access, like, existing repositories and things like that, but a lot of the work that I do is oftentimes, like, Knowledge work and things like that. Um, so it was nothing that was in, like, an existing repo, and I couldn't just keep working, but I've seen people use, like, Claude, C-L-A-W-E-D, where they're, you know, working with Claude Code and WhatsApp and all kinds of other tools, but I don't know. I feel like I already spend enough time with Claude Code. I don't really need it everywhere I'm chatting.

AI assessment note: “I don't. Uh, I tried to do it one time.”

Q AI feature is really interesting. Impressed that it worked too. I mean, my real question with 4.5 is like, okay, great. Now it's made a landing page, but can it create a SaaS app? You know what I mean? Like, yeah, ultimately I'm interested in pushing 4.5 to the limits and I would, you know, I would be curious if it actually could build an app that I can sell.

A It, it absolutely could. Um, I mean, you know, if you wanted to build this entire product in Claude code with 4.5, you, you could definitely do that. Um, you know, like I, like with, with Claude 4.5 sonnet, like I've taken, um, you know, uh, idea for a SAS from, from zero to production ready. So you, you can certainly do it, um, with either of these models. Uh, now. What I have personally found, and to be clear, I haven't tried to build a full on application with Gemini three yet, but, um, I think that Claude's earlier models, like I would always go back to them. Like when I was, I was trying to use codex and GPT five before 5.1 came out and, uh, you know, Gemini 2.5 pro and stuff like that. And while the logic is good, And the, the thinking and the debugging is good. Um, getting from zero to fully ready was easier with the Claude models. Uh, that, that's just been my experience. So I always kind of try to lean on Claude as the workhorse and then have these other models doing what they're good at as sort of a advisory model, if you, if you will. So here they are like side by side. Now, I mean, if we wanted to, we could, uh, we could see how each of these builds out beyond the landing page and just say, Hey, create a clickable prototype, uh, for, for the concept. And we can kind of get a gauge on how, how it's thinking there.

AI assessment note: “It, it absolutely could. Um, I mean, you know, if you wanted to build”

Q So this isn't just, um, an LLM, you know, making stuff up, right? It's actually true. This is, this makes sense.

A Yeah, this is, uh, this is not just an LLM making it up because what we're doing with this MCP integration to the workflows and in it in is we're pumping all of the right contexts into Claude here, you know, so basing all of this around real data, uh, that's out there. So there's no hallucination. It's not just making things up around, you know, what they might think, uh, people are interested in. It's actually Understanding the demand. So it's finding the keywords. And then we selected a group of keywords and we're like, Hey, go out there and read all of the ranking articles, posts, and websites on the front two pages of Google and tell me exactly what's missing. So this gives you like such a targeted spear. If you're a newsletter platform to know exactly how you can go and attack the market.

AI assessment note: “this is not just an LLM making it up because what we're doing”

Q Yeah. Why would you use your hands like a simpleton?

A Exactly. Once I started using a whisper flow, I feel like I was in the stone age before. So this is going to take a little bit of time to go out and do some research. You can see that, uh, you know, it'll use the, uh, the MCP here with a prompt to go out and find out, like, who are they targeting? Uh, what's the landscape look like? Who are the main players? All of that good stuff. Um, so we'll get that data into, uh, into our system here, and Cloud Code can use that for context. Again, I'll spend like an hour sometimes doing research with perplexity, Crawling websites, getting as much data as I possibly can about my market.

AI assessment note: “Exactly. Once I started using a whisper flow, I feel like I was in”

Q Anecdotally, it does feel like we've, with Opus, 4.5 and Gemini three, it does feel like we've entered a new stratosphere of vibe coding. Would you agree?

A I would definitely agree. Like, I think that the lines are just blurring across the board. Like, you know, the line between a non-technical person being able to code blurred, and I, I would always feel very frustrated with, like, the design portion of, of this whole process, you know. Um, yeah, you can build stuff, but everyone can tell it looks vibe-coded, or it just doesn't look right. Now that's That line is getting blurred. So all these lines are just blurring and, you know, it's not necessarily like vibe coding or vibe marketing or vibe designing or whatever It's just like vibe building in a way. And the entire stack is just kind of coming together into one place, which is nuts. All right. So there's Claude. Let's check in on, uh, on Gemini and see what we've, uh, what we've got there. All right. Here's Gemini's version.

AI assessment note: “I would definitely agree. Like, I think that the lines are just blurring”

Q So what do you do when you see errors like that?

A Uh, fix error. How can you make this work? No, um, you know, for, for me, it's, yeah, just kind of a process of, like, trying to dig deeper into what the root cause is. So I use term, the terminal for all of the work that, that I'm doing. So if we think back to that cursor setup, uh, I may say, hey, Claude, uh, ultra think about this, uh, problem. You know, if you type in ultra think, it'll change colors, and it'll use more, uh, you know, power to kind of help to dig in or solve the problem. Other tactics that I've used, um, with Claude, because that's like my daily workhorse, is I'll say, hey, if I'm really stuck, I'll say, hey, spin up specialized sub-agents that are gonna go and find the root cause of why this thing is breaking. Um, I need, like, a Q&A tester. I need a senior sort of, like, Engineer focus on this or whatever, and then it'll spin up these specialized agents that'll identify different areas that they're going to go and research in the code base, and then it'll come together with like a comprehensive sort of analysis for me, and then I can say, hey, all right, cool. Plan this in terms of like a step-by-step process that we can go through to fix it. If it's still stuck, that's where I'll use cursor agent with, you know, GPT or Gemini or something like that for kind of an outside perspective. Now, One of the, the things that Anthropic is kind of talking about wit…

AI assessment note: “just kind of a process of, like, trying to dig deeper into what the root cause is.”

Q And do you need to like install Remotion?

A You do. You need to install it. Uh, again, if you, if you go to x, um, Remotion posted this video, uh, that has their, uh, it has like a free GitHub repository of how they built the video, uh, and has a video there. You can see kind of how they did it. You can get that link to the repo. You can throw it in here and, uh, it'll install all the packages for you. And, uh, and you can essentially vibe create videos just how you would anything else, um, which is pretty cool. All right. Um, let's see here. So we're setting up a page and you know what? I'm going to go ahead and throw this command in here for remotion. I had it in my other window from doing some things earlier.

AI assessment note: “You do. You need to install it.”

Q Just as you're clicking around this, it's like, why doesn't this exist?

A Yeah, it's a really good idea. Couple pages, uh, aren't built, but, uh, yeah, you get the idea. You can post an update, sharing the news with your family. This is a great idea. It kind of reminds me of like, um, What was, what's that platform like Carta or whatever, like for your company, like fundraising and financials and stuff like that. There's also been a couple like family office as a product startups that have come out as well, kind of keeping all of that information safe and stuff. So there's some, there's some good sort of examples out there to pull from for this. The only thing is like for this, I would say, What's the recurring kind of, kind of use case, I guess? Like, one way to think about it would be, um, ok, the estate clears. What, what's next in the journey? Right? Um, do, do you need financial advisory or financial help? Um, do you have some sort of like estate law play or integration that comes next? So like, this is one step in the journey. Um, I, I would say to really improve the concept, it's like, what's the recurring use case after? That, that's the next phase of life after this.

AI assessment note: “What's the recurring kind of, kind of use case, I guess?”

Q And, and in terms of like, how big could this be? Like Nashville car wrapping, like, you know, are we talking 50,000 dollars a year of profit, 200,000 dollars a year of profit? Like, what are you thinking?

A I mean, I can speak to like, uh, to the mobile diesel mechanic, uh, business in Charlotte and Columbia, right? If you think about that, uh, you know, top of your mind is probably, it probably doesn't sound like that big of an opportunity. Uh, however, um, average customer size currently is 1600 dollars per job. Um, we get probably 10 inbound phone calls every single day. We can't even service all of the calls that we're getting right now. We're trying to hire Up to keep up with the demand. So, um, you know, I think that this alone could be like a multiple seven figure business in two, uh, locations. So once we figure out those two locations and we have the playbook and we can scale that out or whatever, um, you know, it can be really, really significant. Um, again, uh, I'm going the route where I'm actually building out this infrastructure and the operations and hiring the people, but you don't have to do that. If I were doing it the way that we're talking about with more of a lead gen play or a lead gen model, uh, there's really kind of two ways that you can approach that. Uh, one is, you know, you create that newsletter, uh, type thing and, uh, you, you acquire the end customers and you go to the business and you say, hey, like, um, I'm gonna, uh, sell you leads for 100 dollars, 200 dollars a piece or something like that. You know, if a company's got like a 1600 dollar averag…

AI assessment note: “I think that this alone could be like a multiple seven figure business”

Q And do you often have two instances? Is that your kind of go-to?

A Sometimes I do. Uh, you know, I, I kind of did that here because perplexity was taking some time and I wanted to keep moving. Uh, I'll do that at times. Like you could have one focus on research and using MCPs. Those can fill up like context pretty quickly. So sometimes it helps to keep that in like a split terminal and then do other work, uh, side by side. So you can like save research, save files here that you're Gathering from like the, the third party tools, and then you can use like your skills and sort of execution in a, in a separate terminal. So, um, I trained this direct response copy skill on like the greats of copy and direct response. So, you know, I looked at like Eugene Schwartz and Claude Hopkins and all of these classic copywriters. And then I sort of updated it for like, you know, today's world, today's internet and, uh, made it somewhat elevated so that it captures kind of like, you know, a solid, uh, a solid brand voice. So, it generates, like, great copywriting, I found. I've used it for all of my websites, and, uh, it, it produces some really good results. So, okay, like, we have some copy here. We're just gonna keep going, uh, in the interest of, of time, but, um, let's see. Use the Playwright MCP to find an agency we would compete with. Capture some screenshots of their website.

AI assessment note: “Sometimes I do. Uh, you know, I, I kind of did that here”

Q Which I think a lot of people miss, right?

A Yeah, a lot of people miss that part. They hop in and they just start trying to create stuff, and that's like where the AI slop thing comes from, you know. They just go into a tool, they throw a prompt in, they get an output, they throw it out there, and then they wonder, like, why, why am I not getting good results? They're skipping kind of the boring research step, uh, oftentimes. Alright, so, uh, normally I would wait, uh, for this to, like, go out and load, but, um, In the meantime, uh, we're just gonna fire up, like, another instance here, and, uh, we'll keep working as that's doing some research. Um, I'm gonna enter in a prompt, and we're gonna assume we have that research, we can come back to it, and I'm gonna start using some skills. Okay, I'm starting a AI marketing agency targeting boring local businesses that do two to ten million dollars a year in revenue. I feel like they're overlooked. I want you to use the positioning angles skill to find out how we can optimally position our company. Now, for those wondering about skills, um, you can create your own skills. Like you can say, hey, Claude code, I want to create a skill around, you know, finding an angle for my company, just like I did. Um, now what I do is do a very deep research process, uh, using perplexity, using fire crawl, using all these data sources to build my skills. I include, like, world-class reference…

AI assessment note: “Yeah, a lot of people miss that part. They hop in and they just start”