How AI Changes 300M Jobs and What To Do (Copy this for 2025)

On Parker, I led tech for a startup that did $73 million at our peak, $47 million at our sale for $23 million. That story, a lot of drama, a lot of learnings. It's for another video. On this channel, it is a daily upload style channel where I teach everything that I'm learning as I go and build a $100,000 a month in profit AI services business. And I cover a couple different things. And let's just get into it. So that's the hook, right? We go through some questions and answers that are sourced from the comments as well as in troublefree. Troublefree is a group of 150 aspiring or current business owners who are leveraging AI every day and our techno optimists. So we had a call today where we went through a bunch of questions. It was me and Aiden. He's awesome, very, very smart. And it's really cool because we got to share knowledge. One person is not as good as a bunch of people, right? So, we're going to go through questions. We're going to go through some news related stuff. Let me make sure this pen works. I'm going to cry if it doesn't. Yeah, it does. Cool. I re-recorded this because for some reason it didn't work the first time. Of course, going to cover the news. Only news that matters if it has a business application or if I have some sort of take on it. You can find AI news literally everywhere. It's spam at this point and it's typically sensationalized because they have to hit a deadline. So, I only talk about if I think it's important. We'll go through some strategy and then we'll go through a build plan. If you want to actually watch me build one of these automations that I plan to sell or use for myself to get leverage, then you can check out Parker Rex Builds. It's another channel I just started that I just record and narrate stuff as I'm actually doing it. And these videos are usually about 45 minutes. It gives me 15 minutes to do the post-production chop chop chopping and uploading. So, it takes 1 hour total. So, first off, this OpenAI clip, I don't have it pulled up, but basically it goes through the comparison of OpenAI versus DeepSseek in these open source models. OpenAI, I don't believe this is actually accurate, but even if you bulk it that it cost $50 billion to make GBT4 and it took Deepseek6 million reportedly, even if you put a 10x multiple on that six and it was 60 million. Well, Deepseek was like, let's call it, it's maybe 5% off the current soda. Soda is state of the art. You'll see this literally everywhere. Even if it's 5% off what this is capable of doing, guess what? It's like 99% cheaper. So, it's something to think about. And endgame for this is they'll ultimately run on your iPhone. That's right. Everyone is asleep at the wheel thinking that Apple is an idiot. No, they're just not lighting money on fire. They're letting other people do that. You see other major macro trends where Satia from Microsoft like $3 trillion company, they buy 51% share in OpenAI. Now they're saying that models are commodities. OpenAI is only valuable because of the consumer play they're doing. So something to think about. Vibe marketing thread. There is a Vibe marketing thread. Everyone knows of Vibe coding. I'm just going to go through this news and then I'll do the questions, but everyone has heard of Vibe Coding. This guy Greg posted about Vibe Marketing. I told him, "Stop talking about it. You're ruining the alpha for people to print." Meaning people don't know about this stuff yet. This is not even close to reaching the masses. But similar to Vibe Coding, where your tools amplify your skill set, the same thing is happening with Vibe Marketing. I think vibe coding requires base knowledge in the subject matter to be good at it. It's a multiple. It's a lever. But if you're bad, if you don't know stuff to begin with, you don't know how to prompt it. People mistake this with someone who puts out a video game like Peter Levels. Well, guess what? He coded for the last 15 years, so he knows how to prompt it better. The guy who coined the term is the best coder on the planet. Like, he's like a top 0001enter. you're not him. You're not that guy pal. So with Vibe marketing, this is that similar concept of it puts a multiple on the existing skill set. And what I like to do is I like to think of what is the future in three, five, and sevenyear chunks. So in the case of marketing, I think if you went back in time to get historical concept uh context, you first had auto posters. So social media really hits the scene in like like 2007 to90 09 that's social network right that is the movie the Facebook movie Zuck makes the first version starts blowing up people realize inbound marketing is a thing Gary Vee starts calling the shots pay attention to Gary Vee he's annoying but he's right people start making auto posting that becomes the trend oh we can auto stuff. That's cool. That goes out of style pretty quickly because as soon as the platform gets large enough, they say, "No, we want you on the platform to schedule these." That's 2010. Has a good run. Next big trend, people realize they can do automations and sequences with trip campaign marketing. The next one is copywriting with GPT. They realize they can train it. They realize they can get it to write some stuff for them. And then now we're in this phase with automations. So with automations you can go and you can have all these fancy innate end flows. So they allow you to have omni channel meaning many channel channel being YouTube channel being Pinterest, Tik Tok, uh blogs, IG, whatever. It allows you to break them out. So I might have like what I'm doing with my content system. I'll have a long piece of format that starts here and then it'll branch out and have all these conditionals based on where I want it to go, what I want it to read like, what I want it to be trained on, all these things. And then you can have it flow back into how much engagements it's getting. So you could add that to the training. But the way that I see this vibe coding thing going is right now you have people that are like the inade boys of the world, right? And there's a lot of people online teaching how to make these. they don't make money off of them or the companies that get them, it's like a bar trick. I'd say that's probably 90% of the stuff out there. It's like this dog and pony show where it's like, "Oh my gosh, we need this thing. Look how cool it is." And then there's a bunch of info product people that do that. But there is 10% who are crushing it. It's the same thing with like drop shipping in the early days. But what I think is going to happen is over time if you had a team like in this case let's take Dolingo Dolingo crushes it on social but they have I think a 10 or 12 person team that's running that and I think what's going to happen I'm calling it now is this will be the future where the role of let's say this was your Tik Tok expert you may have like three people on TikTok. You have a strategist. You have the assistant that does the posting and the commenting. You might have some people offshore that are going and doing the engagement because the action happens in the comments. You might have them going on Reddit to find hooks because that's where people go to get hooks. But this is going to all change where these people are going to start leveraging parts of these big flows. So if this was Tik Tok and they had four steps or two steps in the conditional, they might be 10 steps. They might be bringing in like, oh, if this type of content made it not go well, let's bring it back in. So it'll be incredibly data centric, data informed, not data, because if you're only data, you can miss out on anomalies. That's something to think about when you're working with clients. for content. If you have the weeks across the year, do an 8week period and you have your number of impressions or whatever it is your your KPI. If you just did data uh driven and you're like this one thing we do works and when we did it, it went up on the second and a half week, let's keep doing that. Well, it's slowly going to come out of style, slowly going to stop working. So, you want to be informed on it, but you also want to take shots. And you take shots by having leverage. You get leverage by doing this. So, this is where it's going 100%. And it's funny, I put Pinterest in here in the first time that I recorded this because it's a lesser known one, but it still gets billions of people in traffic and it's not leveraged up. You can start adding more and more of those channels. As you have the leverage, I cover this in my content pipeline that I'm building out on the ParkerX builds channel. We're on step one of many, but you basically start by listing out all the different pieces in the system so that you can take a piece of long form content like this. Going to have it transcribed for you. This is part of the editing process for me. This is like where I spend the 15 minutes after the video so I can do it in an hour where I upload it to Premiere. It chops any filler words like um or uh and then it goes through this whole thing. So maybe there's eight steps for you to have it go from one video to 35 clips across Instagram, across YouTube reels, across, sorry, YouTube shorts and Tik Tok. And then you get into the text and the voice stuff, right? So it's going to go and write a master guide on step-by-step instructions SEO optimized for blog. Then it goes and posts. You can drop HITL, which is human in the loop, into each one of these steps. So you can have a stop gap where it's like, "Nope, I need to check it. Nope, I need to check it before it goes to the next." And then it's stack ranked based on the market share of each, right? So if you go and you see what's the search volume, everyone thinks Google's like the only one, but guess what? There's something underneath that called domain authority, page authority, domain rank. These are the scores. These are three of 200 of them that Google give you. And you need to just nail it on all these. There's a reason why people that are all over the internet show up in chat GPT responses. They're not reinventing the wheel on search results. So that's where it's going undoubtedly. And people like you, people like me, those who are technical, who have good fundamentals can apply these automations and leverage the hell out of them. So that's what I'm going to be doing. So that is that rant. What else did I have on this? Yeah, I kind of just think that this will be table stakes for marketing teams. Right now, it's this new flashy thing, but it will be what's called a table stake feature in the future. Table stake is a Silicon Valley term. If you're building out Venmo, a competitor of Venmo, there'd be things that you'd expect to even get off the ground. This is a poker analogy. It's table stakes. To be able to be at the table, you need to have these things. So, I believe NAN will be a table stake in every marketing team. It means that of the 300 million jobs, so 300 million jobs displaced by 2030, right? This is a soft stat. I think it's going to be way bigger than that. It's going to be these jobs. It's going to be coding jobs. It's going to be every type of job. And it's your job to take advantage of this stuff. Get on the trend. Jump on it. Learn as much as possible. Don't buy from info product boys unless they're doing the thing that you want to do. Cool. Let's get into some questions and answers. I Let's see. Let's go in here. The first one that came in, I already responded to it. The first one that came in was, let's see, nice video. I have a question. Do you feel like you connect deeply to both or either of these two niches? or do you feel like there's another reason besides the act of serving them in itself that is actually the main driver for you deciding to serve these audiences? Thanks for the question, Brian. I answered this before, but I'm going to just kind of review the notes that I had. So, I've made uh a lot of different YouTube channels since I was 14. I'm 32 now. My birthday was a couple days. You can probably see the balloon there. But that's 18 years of of trying and learning and trying and learning. The most successful one was called Parker the Pyro and I made a channel basically teaching people about chemistry. That's when I was 13 to 16. I got kicked off. That's for another video. But what I learned then was the success of it was predicated on me really being passionate about the thing I was already doing and then there just happening to be a camera there. I liked teaching people. I like to learn something on the internet from very niche uh PDFs and forums and books that I'd order online and I want to package that up so I could just share it which is what I liked doing. So I did it and that channel grew to about 14,000 subscribers and that's in 2009. So it had success because I just pumped content and it was authentic and I liked doing it. So, I tied passion to it. And what I learned next, and the next uh channel that had decent success, got to 5K, is this main channel. This main channel is just Parker Rex. What I did was I filmed something I was really passionate about at the time, which was working on this startup called Delivery Dudes. And the type of content that I'd make made the job more interesting. So, I'm working every day at this startup. I had been there since it was literally nothing, like zero dollars. And I had been responsible in helping it get to 47 then 73 and then after 73 when we went back to 47 this area it really it it it sucked like I wanted to leave so many times I found myself at Y cominator pitching there found myself trying to start a drop shipping store selling coffee filters off a coffee theme page. I was trying all sorts of stuff but what made it interesting again was this addition of the YouTube and I just started teaching again. I started making videos about product management, which is a very niche job. It only exists in large tech organizations who don't have the CEO leading the product strategy anymore. You don't know about the role. You basically figure out what you're going to build and why you're going to build it. And it's fun because you get to work with designers, you get to work with developers, you're quasi technical, you're responsible for what gets shipped. And that matters a lot at an organization at our scale where you have 5,000 restaurants you're working with. You have thousands and thousands of 1099 contractors who are using your app every day to perform the deliveries of the food. You have, you know, 500,000 monthly active users using the consumer products. So, it's a lot of fun. It's a lot of responsibility. So, I wanted to teach about it because I think it's this niche role where if you didn't go to, you know, business school or you didn't go to Stanford or something, then you probably aren't going to get into it. But I wanted to disprove that. Basically, distill everything that I was learning. I made vlog style content about us launching something during um the pandemic and it was fun, but as soon as it wasn't fun, I stopped making the videos. It felt like a job. And so the trick is to find the thing that you enjoy talking about that you could talk about all day and make content around that. Now, why I'm splitting them out is intentional because what I learned from the product management audience that I built was I ended it by getting about 5,000 people and most of the views were on videos that had nothing to do with product management. I had 100,000 views on a video about a design tool. So, I guess that's in the realm. I had close to 100,000 views on the video about um like what app to use for working out. So that's like a technical review, but it was more like the big views that I saw were really really wide, right? Wide audience. So it's like the top of funnel was just like software reviews. It was about the Whoop app, Fitbot app, and about Envision at the time, which is gone. And then the product management was towards the bottom. It was like not even towards the bottom, but it was just a skinnier funnel. These are separate funnels. So then you had like really small. Let's make a small funnel. This is PM because it's tiny. What was interesting about this though is of the 5,000 I launched a product that was a $200 info product that took everything that I learned going from 0 to 73 million and distilled it into like 20 videos. And it was really easy for me to make because I just talk. I made structure to it. Tried to add some production value. But then you go and you sell $20,000 or in I think it was like 18.3 or somewhere around that. But I did that in about two weeks. And then I felt scummy. So I was like, "No, I don't want to do this." And I I shut it down. Stopped making the content. I shut down the info product. But I learned. And so now this channel, I'm realizing it serves a different thing that I'm really interested in. I treat it like I don't have an audience at all. I started at zero. Even though I have 5,000, it's a new audience because a lot of those people that were attracted, the magnet was product management. Maybe they didn't go into product management. It's been 3, four years. Maybe they already have a job in it. They don't need newbie advice. So, you kind of learn through sets and reps of like how do you match what you already like to people are on a trend to do. So, if you wanted a big channel, talking about AI in general for non-technical people is going to pop. It's going to go to a million subscribers. I'm I'm guessing 18 months from now this channel will be up there because I'll be covering very large topics that follow a large trend that I already like talking about because I have to do all these different jobs of uh building a client services business and having a SAS that I need to hire people for when I have the funds to do so and running a research lab community that is 150 people a business owner. versus aspiring business owners that leverage AI and believe in it because I do all these things. You have to learn all these skills. So, I'm learning how to leverage AI for sales, how to leverage AI for marketing, how to leverage it for recruiting, how to leverage it for design, how to leverage it within marketing for copy. Copy has emails. Like there's all these like sub things, but they still go under this umbrella of non-technical AI. And so, this is a wider funnel. It's going to make me get more uh more eyeballs around the other areas of the internet that I'm on. So, someone that is technical might find this because they're a founder and they see it. And ultimately, this will draw more attention towards the events that we do for troublefree, which means at the end of the day, it's a marketing funnel. I don't plan to sell any sort of info info product. I think I just have such a I don't like them. I just don't. So then you have the daily channel which is behind the scenes. This is something I'm already going to do. It's something I'm passionate about. I try to get in the habit. I've been trying to get in the habit of writing down the ideas that I have and then going through them. Not just like scribbling them in an Apple note, but literally taking pen to paper or pen to tablet in this case and going through what are the ideas that I have? What's the strategy I need? what's the news that I'm going to need to apply to move the needle forward? And then what do I need to build in order to to accomplish all of my goals? So, by just doing it already, I could just add questions and answers into this and then hopefully drive a lot of value for people that watch, give them a behind the scenes, get them to be inspired to go do the thing without having to go as much like through as much hell as I've had to go through and then ultimately build a community of these business owners, aspiring business owners to go into one of my SKUs, which is this bootstrapped research lab. I'll go into the SKUs in a second because it's not just like, "Hey, join my thing. Give me $39 a month." It's, "Hey, join my thing. It's now our thing. How do we make it better? How do we push each other?" We had a call. We had the first call today and Aiden, he's my boy now. He showed up and I got to learn a whole bunch of stuff and we pushed each other. It's this intellectual sparring that you do. And I love that. And it expands your your reach. Like if I'm here and I'm in Florida, which is like the the wiener of of the US, like look at that, and he's up here in another place, not Florida, there's no way we're going to chat. The same thing goes with I had a co-founder in this dumb idea that we did. It's not his fault. It was my idea, but he's in Cape Town, South Africa. How are we going to connect? Well, YouTube. That's how we met. It's like Tinder, right? It's a it's a a dream story. But no, we met because he made videos about Flutter, which we ended up hiring him for. So that I've been the customer. I've been the person that hired someone. He made hundreds of thousands of dollars out of some Floridaian that's over here yapping and watching his videos and being like, "Okay, cool. I trust this guy. Like, can you do this? Here's $10,000 for the first month." So, it deepens that connection. And then that type of person that it's going to attract is a business owner, aspiring business owner. Maybe people want to learn how to plan agents, all that sort of stuff. Maybe they want product strategy. So again, very intentional. And then the last one's the Rex Builds channel because people that are going to watch this, they are technical. They want to get into the nitty-gritty. The way that I speak about it and the way that ultimately I'll get better at talking to different audiences is you get to change the way you communicate. So, I wouldn't use phrases in the main channel such as use rag, connect that up to a vector pipeline, then use superbase, then use pg vector. PG vector does this, pg vector can do a trigger to that, like all these jarens. Jargon's fine if you're talking to the right person because it saves a lot of words, but it's not going to be helpful on that other place. Yeah. Yeah. So, that's Basil and she's confused. So, I can just basically make Basil content for the first one. It's going to get a massive audience. It's going to get more eyeballs, which means more surface area on the internet. That increases your odds of luck, which ultimately means more people get helped, more people learn, and you're kind of like reskilling, which was one of my big goals when I started out with this trouble-free uh AI idea. So yeah, that'll bring in coders, designers, or aspiring coders, aspiring designers, also product managers, and then yeah, you end up on a call with someone from somewhere else that has literally the same exact interest as you, and it's just the best. That plays into the SKs and incentives. Always follow the money, right? You never ask a barber for a haircut. That's Charlie Mer. And it just makes sense. He's always going to say, "You need a haircut." Because he makes money. So the way that I make money is actually off right now it's primarily AI services. I had done SAS for a long time and I made no money off of it. I made 18,ou no I think it was 18 to 22,000 in market revenue off of the um Airbnb for music thing that I did with the guy from South Africa. But that wasn't money in my pocket. That was just like horrible unit economics. So then I said I'm going to do SAS and in order to do SAS I needed to learn AI. In order to do AI I needed to know how to code. There was no vibe coding three years ago. So I went all in. I had a basic understanding of the front end stack. But then I said I need to know back end. I need to know next. I need to know React 19 and 18 at the time and next 14 at the time in order to unlock this. So in order to unlock this I needed the cash to cover. So AI services with SAS. My SAS app called MAP has the opportunity to be a billion-dollar plus business because I think it's going to impact a lot of people. It's a multi-aggentic platform that helps you become the best version of yourself. You have a model. It's trained on all of your goals. Breaks that out into different products, different SKs, but it's multi-aggentic. It's a pain in the butt to build. You needed more time. And even the opportunity for a billion dollar exit or a billion dollar valuation is maybe 5% if you're awesome. So then I say, "Okay, AI services goal is to get to 100K profit." I got about a quarter of the way there, but it was profit that I did not have leverage on, meaning I didn't have more hours in the day to do stuff like this to build a community. So I said, I'll take half of that, but it's going to be way higher leverage. Let me get really good at this new offer. Let me strategize a new offer such that I can get to 100K in profit while still going to the gym, while still having a relationship, while still doing the daily uploads and all these things that I need to do where I'm not going to want to like tear my hair out. That brings in the third skew, which is the bootstrapped research lab. This is called TroubleFree, and it's $39 a month. The point of it is not to make a living off of this. The point of it is one, it's a filter. Two, in order to make something good, you have to water it. It's like a plant. You need to invest back into it. So, by charging $39 a month to aspiring or current business owners, it puts a filter. And you take the first 2925 and then you ramp the price up for the second half of people to 100. You end up with a little over 10k a month. And then that 10k a month, it goes to work on building out the assets to improve the value of the brand, the content of the brand, the overall surface area that that has because trouble free will be its own standalone thing. It's not the Parker X guru. I don't know more like me alone here just yapping at the internet on my experiences is nowhere near the amount if we just had like 150 people, right, that are qualified. And then you put rules in place like we have standard operating procedures that we're going to have where it's like if you do put a prompt in there. This is not a little prompt uh treasure chest where it's like, "Hey everybody, here's my video. I'm going to make this thing. Uh join the school and you get that." No, I hate that. Instead, you can only have a prompt in there if it's verified. So then what's verified look like? Okay. Well, do you have receipts where this has made you money? Do you have evals in a Google sheet that we can look at that you've you basically have context of why this thing matters? So you know if 149 other people are looking at this that there's proof probably like expiration dates on it. This is just one example, right, of like how you'd put the money to work. You take the money, you put it back into content strategy, content assistance, automation buildouts, and ultimately building your own platform. Because I've got my problems with school. I think it's got really a big brand issue because there's so many scammers and I mean any new trend people hop in and start slinging product out of it. But everything from like the way that the content is displayed to the way that the chat works, the tech that's used, it just feels like it's from 1995. So then that'll go into be building its own platform. that platform will have automations around it to basically own the internet for certain longtail keywords around communities and it's just like awesome to be able to talk to like I had a I was part of a group when we were at delivery dudes and we would go and be part of this group our executive team a couple members would go to it and I remember like just when you go through something that's really hard in business and you're meeting with someone that is also like at your same scale and you can talk to them about it. That pure group is sick and those cost a lot of money. So that's that question. Thank you for the question. That's the point around all these. The next question was about from the itchy chicken. If you're good at AI, there are ways to fill the gap without the basic knowledge you're talking about. It takes advanced and in-depth knowledge of how to use and guide AI, which I've seen few people do, but it's definitely possible. Hard to teach comes with hundreds of hours of experience. So, uh, itchy chicken, you got a bad case of copium. I'm just going to say it. It's It's mostly a joke, but it's, you know, it's a funny meme, too. What do I mean by that? I think that if you had those hours allocated towards learning the fundamentals, you're better off. I think that it's similar to like Web Flow. Web flow is a tool that was awesome for a couple years until now it's like piece of junk and totally unusable because you can move quicker doing it otherwise and that skill set it's like a quasi skill set. It's between HTML CSS and design and it's weird. I think that if you have a couple options option one is hundreds of hours learning clue GI tricks or option two is learning the basics of programming. These have more staying power and they have more leverage. The reason that Andre Carpathy, the best people in the world, are able to vibe code is because they know exactly what to say. They say, "Go into this file, use this library, make sure that you hook this to that, don't do this." Those prompts are very informed and they're based in knowledge. They're based in experience. And you learn how to do it by eating crap, by messing up, by banging your head against the wall. You're going to do that anyways over here. And I I did on the first recording realized that like you've probably seen people like McKay Wrigley, awesome guy. I think it's great that he put out these demos. He's gotten a ton of clout and attention. And at the end of the day, I don't care what you do. I just think that this is not the way to go. Like the charts are going up and right, right? If you see the number of Superbase accounts, they're going like way up here. And I would assume that like everything in here is the newbies. Like that exponential growth are the newbies. And I think that it's important that they go and they have this easy on-ramp by doing this, but then they'll end up whether it's guy doing call it twice as much of this, but they're willing to put up with just spamming and burning tokens versus half as much of this of just knowing what they're doing. As long as they're doing it, it doesn't matter to me. It doesn't affect me. I state that vibe coding is a hund00 million industry for those that clean it up because it's just going to be insane. Like this is a uh a dream come true for security folks because there's going to be so many security issues and um at the end of the day I don't care. But I do think that this option is much better. JavaScript and Typescript are not going anywhere in 10 years. Sorry. I had someone say that we're not going to need Python. It's going to go away. We just need ones and zeros. I say, "Hey, that's probably, you know, I'm thinking that sounds like something that would be said at a cocktail party where if you're talking to someone non-technical, they nod their head. They say, "Yeah, that's that's sure something." Sure, sounds good. It's bold, but it's not true because if you look at the way that things work, well, guess what? Yeah, Python's a wrapper around C. Uh, Apple's a wrapper around TSMC. Like, how far do you want to go? Stripe's a wrapper around banks. Like, point is is all these things connect to one another. There's a zillion dependencies. Python has a massive ecosystem. Most of the biggest companies are built on these packages, whether it's in one ecosystem or another. So, knowing these things won't go away. If they did, if I steal it, I said, you know what, they are going away. Well, guess what? the skills aren't. So, your prompting skills, they're going to be way better. You're going to know what to do, what to say, how to say it. Your debugging skills are going to be better. Your ability to think as a systems thinker, maybe the most important one. That skill set alone, that's the alpha is like, how do I think in systems from p first principles? How do I do that? Because it'll apply in other areas of business. she'll start thinking of everything as like a a theory of well basically like a like a pipeline like every every company is kind of a pipeline and so that's that's just how I I feel about it. I wouldn't dedicate the time and energy into maxing on this. You're going to do most of it anyways. Like I still, you know, having done this and having coded a bunch, you know, you put the 10,000 hours and then put a like that's just the starting line. like you need more than 10,000. 10,000 hours is achievable if you're working doubles like two years I think it is. Um but please go this direction. Stop huffing on the copium. You'll be happier. And the way that you do that is you just build a bunch of projects. Like I was showing Aiden today, like I have a folder just filled with projects and then you know you get like a hundred of them in there. These little things that you're building. Oh, I needed to build uh what did I need to build? I need to learn how to use a table with a new library. Okay, I need to build something that pulls in tables and does a a real-time calc on how much obesity costs for each country globally. Okay, well that taught me a bunch of stuff. That's easy. So, you can learn by doing. You don't have to sit there and read the docs. You can also use a doc rocker, meaning like you have an agent that goes and does a fire crawl on something. You don't know how to do that? Well, you're going to learn HTTP requests. Build your own curriculum and then build the products around it to learn faster. It makes it more fun. That's how I did it. Thanks for the question and uh ultimately take your own path, but that's just one man's opinion. Let's see. Really enjoying this seeing the narration of your work and how you're building AI first. Thank you, Vvec. I covered this, but what I wanted to say is like don't automate anything you tell you've done it a bunch of times. So everything that I do is a list. Everything everyone does is a list. And so then some things aren't automatable. So, if I have something that's simple and I put a little green check next to the thing that I again, I've done all these 10 times, I might have like 75% of them that I can automate, but then this one put a little red on it. I'd say, nope, that's hittl or human in the loop. And then the way that you think about this is when I was first starting to study this and take it seriously, I was doing a bunch of searching on prompts and how memory works with this and chain of chain of thought. All these terms, well, they come from the MCAT prep. I started coming in to pages that were showing MCAT, which is the medical prep course. And the reason that is is because if you watch old talks of um what's his name? Andre Karpathy or the other guy who started SSI that recently quit, but he was like the chief scientist. They talk and they almost sound like doctors because everything that we've done to shape the way that these systems perform was based and influenced by the way that our brains work. Now, you could argue that the transformer is more like a slot machine, but it's I mean it's kind of just like a uh statistics experiment, but that's the way I think of it. And so whenever I'm doing any sort of workflow whether it's written in code or now in an aden it's just a list and then that applies to your business. So with a business pipeline, whatever your business is out there, they either sell a product or a service, right? And Elon's the best at this in the world of using frameworks to kind of delete steps. Like does a step even need to exist? Like what prevents us from doing that? Question the requirements. Delete the requirements. If you're not adding more in at the end, you made a mistake. But only at the end do you automate. And so, you know, if you had the steps and you hadn't done them a bunch of times, then you might start to automate prematurely and then you start automating all this and you take the time to automate it and it's like, whoa, whoa, whoa. Like, we didn't even need this. The best way to optimize is just to delete. So, that's something to to bring into mind. And then I like to think of every single business now as these steps in the process. So, everyone has their marketing different channels they sales if the marketing didn't work or if it's high ticket and they need to supplement the marketing with the sales they do that but even high ticket stuff if you have really awesome marketing you don't need someone to sit on the phone and make you sweat over it then you have the fulfillment or the delivery obviously SAS is one of the easiest and then you have things on the back end that's like customer success management you can automate some of that but then you'd list out like what are all the steps right so in my case with AI services. It's like, okay, well, marketing, YouTube, and then I'd list out that massive content strategy that I had in the other section. And then for sales, there's an automation for that. Okay. Well, you need proposals. There's like admin involved. There's like follow-up change. Are you doing a 357 where you're following up with people that haven't replied where the SOPs around that? So, you start by just doing it manually. I didn't automate proposal generation until I had already done like four of them. Um, and that's because I had done more of those before and I made a slight tweak on the way I was doing it. So, I had the context to then have the confidence to go and do it. Uh, it just makes it a lot easier when you're trying to set up the systems. So, next up is NAN versus make. I already circled this one. Why? Well, the way that they charge is one big reason. the per execution model is what make does. So you could have like a 10-step flow and they're going to charge you for every single one of them rather than like the whole scenario. So that sucks. Then also self-hosting. I think self-hosting is a big deal. I think that the future of the multi-agentic flows whether in a marketing team, sales team, fulfillment, development, is going to be nodebased because it lets you see the flow of things and you don't necessarily even need to open files. Like when I'm coding and I'm making really good strides, it's me knowing exactly what's going on in the codebase and then seeing the changes and only focusing on those changes, not digging around in random files. So, if I can just see the diffs and I see the strategy and we're on the same page and they're each individual nodes that are really specific and really dialed into that one job, you're really just looking at the nodes and the outputs. And then with the fact that you can self-host, it just makes it better. And I think ultimately you have an agent flow that would deal with provisioning your servers. Then that could be a skew that you offer to somebody later. Someone brought up the idea, actually this was Aiden on the call today about NAN versus code. A lot of coders have god complex. Like, oh, if it's not code, then like so dumb. I fell into this trap for a long time. But save it for the good stuff. Use N if you can. It makes it easy to change. It's visual. So that's kind of nice to be able to see. You don't have to do that extra step of like taking all these prompt chains and then like I don't know like doing a mermaid diagram. I mean, you don't have to use any like gluey cluji rappers. Uh, shout out lang chain, but self host is a big deal. And then I just posted like web flow rip. I had a client that said, "Hey, can you build this?" And I said, "Yeah, it'll be insert big number if we do a custom and then insert a number." I thought that was way too high that they wouldn't take to do it in their existing Web Flow thing. And they said, "We'll do that." And so I did it and I'm like, "Oh my gosh." Like you look up the best person in the space who everyone thinks is the best web flower and then you see them talking and saying this is garbage software don't use it. That's bad. This is a meta skill. This is like what the chicken asked about spending all the hours learning the thing that's not going to be helpful soon. Like non-ransferable skills. It's just not that all the skills are non transansferable. It's just there's less. So you don't have as much transfer which isn't fun. Lastly is the builds. So, at the end of the videos, I just cover stuff I am going to build and that'll be on the Parker X builds channel if you want to go nerd out. I was talking about prompting as a product. I think if you have a really good prompt, a lot of products are just prompts. So, I was using this for a client project where it's like, hey, we need these UI uh designs done. And I was trying to make it into an agent flow where the first one would be just finding the diffs between the two and then this itself could be a product, right? where it's like the UI agent and it's like you go to the uiagent.com you upload the differences and then step one would be to actually find the differences and then step two take that plan have it brought into an architecture guy. Architecture guy gives you the actual implementation plan to do it and then those atomic steps are given to the implementation engineer. They need to be atomic steps meaning like very specific next actions verb noun there's like a there's a format to it. So a bad example would be like build landing page. A good example would be make this component using XYZ very specific no questions and it's uh declarative. Last thing is auto timestamper. This is a tool that I think will just go bananas. So what I noticed was on YouTube videos, Joe Rogan in particular, huge podcaster, right? Largest in the in the world, no one posts timestamps on his videos. Like his editor, Jamie, doesn't put the time stamps. I don't know why. Maybe I should sell him that. But instead, I think a higher leverage thing to do would be to have an automation that posts them there. So it's watching, it sees, it transcribes, it takes the transcription, makes the time codes, posts the time codes. There's my face at the top of every Joe Rogan video until infinity or until he hits me up and says, "Hey, can I have this?" or "Hey, come on the pod." So, I saw this on a video I was watching yesterday with him and the top comment had like 5,000 likes and it's like, "Where's time stamp guy?" And it had like been out for an hour. So, I'm just thinking, wow, how many people saw that? Didn't say anything, but that could have been my face. That could have been exposure. There's a company that does this. is not a company. It is a company, but the idea it's a feature. It shouldn't be a company. No one would pay for it. They'd get like 4,000 people to their site a day. I just don't think it's a a company. But the way that it would work and what I'll build on the channel is it watches for a YouTube video. It gets the channel ID and it's polling. So, it gets the channel ID. It stores the name into Postgress database because I want a front end for this so I can add more channels and see some of the stats. I think that'd be cool. And then it would do a whisper transcription. It has a circle around it because it's the bottleneck for speed. I know whisper just came out with 40 transcribe. So that'll be hopefully more helpful. And then it would trigger using a Postgress trigger to go and generate the script. So now it has a script. It does a trigger to make the time stamps. You could use a superbase edge edge function for that. Once you get the timestamps, put it back in there in the database. And then ultimately, oh, sorry, take the timestamps, that's a variable, and then add, you know, the little tag that's like, hey, if you want to learn how to do this, check me out. And then it'll go comment it using oath 2, and then I'm on every YouTube video for that particular channel. So, that'll be on the build's channel eventually, hopefully later this week, cuz I had to record this twice, unfortunately. Um, was there anything I missed? I believe that's it. Was a good Yap session. If you guys have any questions, please hit me up. I really enjoyed answering these and I hope they were helpful. Uh, Mr. Chicken, I wasn't picking on you at all. I just really like that meme. So, uh, the name of this video is going to be 300 million jobs and how to not get caught or something like that because it is true. You need to scale up. You need to stay on it. Be make it fun though. like this. For me, it's fun. That's why I'm doing this. It's Saturday. It's 5:00 p.m. I could be doing anything, but I'm yapped at a camera. But it ended with me talking to someone that was like-minded. So, if you're interested in being in troublefree, then go and check that out at school.com/troublefree. This is going to be a company in the long run. It's not just uh, you know, me selling you some gated content. And I'll see you in the next video. Do all the subscrib. Thanks for watching.