WEBVTT 00:00:00.000 --> 00:00:04.379 Imagine spending three months building the perfect 00:00:04.379 --> 00:00:07.120 AI template pack. Yeah, just pouring your whole 00:00:07.120 --> 00:00:09.080 soul into it. Right. You finally hit publish, 00:00:09.240 --> 00:00:12.160 you refresh the page, and well, you hear absolute 00:00:12.160 --> 00:00:15.160 silence. It's brutal. But, you know, it happens 00:00:15.160 --> 00:00:17.699 every single day. It really does. Why does that 00:00:17.699 --> 00:00:20.579 happen? Because you asked for their money before 00:00:20.579 --> 00:00:23.760 proving your competence. Exactly. People skip 00:00:23.760 --> 00:00:27.320 the most crucial step. They want to be the expert 00:00:27.320 --> 00:00:30.000 immediately. But nobody trusts a self -proclaimed 00:00:30.000 --> 00:00:32.399 guru anymore. No, definitely not. Those days 00:00:32.399 --> 00:00:35.780 are over. Welcome to the Deep Dive. We are exploring 00:00:35.780 --> 00:00:38.579 a highly practical guide today. It is a really 00:00:38.579 --> 00:00:41.380 fantastic one, too. It tackles building a profitable 00:00:41.380 --> 00:00:44.479 AI side hustle from scratch. And it does this 00:00:44.479 --> 00:00:46.340 through a very counterintuitive method. Yeah, 00:00:46.399 --> 00:00:48.539 it goes against a lot of standard advice. You 00:00:48.539 --> 00:00:51.340 document your messy, clawed AI workflows directly 00:00:51.340 --> 00:00:54.380 on YouTube. You build vital trust first, and 00:00:54.380 --> 00:00:56.740 digital products second. It is a fascinating 00:00:56.740 --> 00:00:59.259 approach to the modern internet. You aren't starting 00:00:59.259 --> 00:01:01.740 with a polished final product. You start with 00:01:01.740 --> 00:01:05.760 a raw, unfiltered learning process. Beat. Let's 00:01:05.760 --> 00:01:08.019 unpack this initial trust barrier. Okay, let's 00:01:08.019 --> 00:01:11.200 do it. The internet is flooded with AI experts 00:01:11.200 --> 00:01:14.340 right now. Traditional side hustles often fail 00:01:14.340 --> 00:01:16.799 at the start. Because nobody knows who you are. 00:01:17.120 --> 00:01:20.180 Right. Strangers simply won't buy your expensive 00:01:20.180 --> 00:01:23.000 AI templates. They definitely won't hire your 00:01:23.000 --> 00:01:25.140 brand new agency. Because they haven't seen your 00:01:25.140 --> 00:01:27.439 technical skills in action yet. It is, you know, 00:01:27.519 --> 00:01:30.120 a classic chicken and egg scenario. You need 00:01:30.120 --> 00:01:32.219 clients to show your work. And you need work 00:01:32.219 --> 00:01:35.000 to get clients. It is incredibly frustrating. 00:01:35.239 --> 00:01:37.260 The source points out a brilliant bypass to this. 00:01:37.400 --> 00:01:39.840 You don't actually need a client to show your 00:01:39.840 --> 00:01:41.739 skill. No, you really don't. You just need a 00:01:41.739 --> 00:01:44.840 problem, an AI, and a screen recorder. It is 00:01:44.840 --> 00:01:46.400 intimidating though, isn't it? Putting your half 00:01:46.400 --> 00:01:48.599 -baked ideas out there. Oh, absolutely. People 00:01:48.599 --> 00:01:50.859 want to hide the struggle. I mean, I still wrestle 00:01:50.859 --> 00:01:53.439 with blank page paralysis myself. Starting a 00:01:53.439 --> 00:01:55.760 new project feels incredibly daunting. Which 00:01:55.760 --> 00:01:58.420 is exactly why this strategy works so well. You 00:01:58.420 --> 00:02:00.819 aren't staring at a blank page. You are starting 00:02:00.819 --> 00:02:05.040 with Claude. Yeah. You use the AI to learn and 00:02:05.040 --> 00:02:08.539 test workflows. Then you simply record that exact 00:02:08.539 --> 00:02:12.060 learning process. But the source emphasizes a 00:02:12.060 --> 00:02:15.060 strict prerequisite here. You must have a clean 00:02:15.060 --> 00:02:18.280 -clawed workspace first. Yes, that is completely 00:02:18.280 --> 00:02:20.120 non -negotiable. What does that actually mean 00:02:20.120 --> 00:02:23.120 in practice? Well, it means setting strict memory 00:02:23.120 --> 00:02:26.000 limits and context rules. If your workspace is 00:02:26.000 --> 00:02:28.900 messy, the AI gets confused. It starts pulling 00:02:28.900 --> 00:02:32.300 in weird data. Exactly. It might reference an 00:02:32.300 --> 00:02:35.219 older, completely unrelated conversation. If 00:02:35.219 --> 00:02:37.560 that happens on screen, your audience gets lost 00:02:37.560 --> 00:02:40.120 immediately. The foundational layer must be crystal 00:02:40.120 --> 00:02:42.840 clear. Right. You want viewers to easily follow 00:02:42.840 --> 00:02:45.840 your internal logic. I love this underlying concept. 00:02:46.909 --> 00:02:49.330 It's like an open kitchen in a restaurant. Oh, 00:02:49.349 --> 00:02:52.210 that is a great analogy. You see the chef prepping 00:02:52.210 --> 00:02:54.430 the ingredients. You watch them adjust the heat 00:02:54.430 --> 00:02:57.490 on the stove. Seeing that process builds your 00:02:57.490 --> 00:03:00.009 overall confidence. You actually trust the final 00:03:00.009 --> 00:03:02.250 plate of food more. Right. And you're showing 00:03:02.250 --> 00:03:05.069 them the raw digital ingredients. Exactly. Let's 00:03:05.069 --> 00:03:07.569 say you take a rough idea about AI side hustles. 00:03:07.870 --> 00:03:10.449 You ask Claude to generate five different YouTube 00:03:10.449 --> 00:03:13.250 video topics. You ask for a hook, a lesson, and 00:03:13.250 --> 00:03:16.180 a clear takeaway. Yeah. Then you show viewers 00:03:16.180 --> 00:03:19.199 exactly what Claude spits out. You review the 00:03:19.199 --> 00:03:21.379 output and explain your editing choices. You 00:03:21.379 --> 00:03:23.520 explain why one hook works better than another. 00:03:23.639 --> 00:03:27.199 Exactly. It is highly educational to watch. But 00:03:27.199 --> 00:03:30.379 here's my question. Does showing the messy middle 00:03:30.379 --> 00:03:33.159 actually build authority? Or does it make you 00:03:33.159 --> 00:03:35.900 look like an amateur? It definitely builds authority. 00:03:36.340 --> 00:03:38.939 It shatters the illusion of AI magic. People 00:03:38.939 --> 00:03:41.840 know AI outputs are rarely perfect on the first 00:03:41.840 --> 00:03:44.300 try. They know it requires tweaking. Right, so 00:03:44.300 --> 00:03:47.219 watching you fix the mistakes proves your real 00:03:47.219 --> 00:03:49.800 competence. It shows you are the skilled operator. 00:03:50.180 --> 00:03:52.840 So breaking the AI illusion actually proves your 00:03:52.840 --> 00:03:55.800 real competence. Yeah, authenticity beats polish 00:03:55.800 --> 00:03:58.659 every single time. So if the goal is radical 00:03:58.659 --> 00:04:01.680 transparency, well, showing people the actual 00:04:01.680 --> 00:04:04.360 messy middle, a text blog probably won't cut 00:04:04.360 --> 00:04:06.379 it. No, not at all. They need to see your screen. 00:04:06.590 --> 00:04:08.930 Is that why YouTube becomes the default engine 00:04:08.930 --> 00:04:11.889 for this strategy? Exactly. YouTube physically 00:04:11.889 --> 00:04:14.530 proves that you did the work. Viewers literally 00:04:14.530 --> 00:04:17.209 watch your cursor move across the screen. They 00:04:17.209 --> 00:04:19.550 see your very first idea. Yeah. They watch you 00:04:19.550 --> 00:04:22.250 type the specific prompt into Claude. Just to 00:04:22.250 --> 00:04:25.269 ground this for a second, a prompt is simply 00:04:25.269 --> 00:04:28.209 a specific instruction telling AI what to do. 00:04:28.430 --> 00:04:30.850 Right, that is exactly it. They see that prompt 00:04:30.850 --> 00:04:34.310 go in, then they see Claude's raw answer appear 00:04:34.310 --> 00:04:36.750 on screen. And it is completely unedited at that 00:04:36.750 --> 00:04:38.990 point. Unedited and usually a little robotic 00:04:38.990 --> 00:04:42.089 at first. Then they watch you make critical edits 00:04:42.089 --> 00:04:45.300 in real time. They even see you design the final 00:04:45.300 --> 00:04:47.879 Canva thumbnail. Right. They see the entire messy 00:04:47.879 --> 00:04:50.399 process end to end. Wait, if Claude does all 00:04:50.399 --> 00:04:52.680 the heavy lifting, though, aren't we just spamming 00:04:52.680 --> 00:04:54.939 the internet with generic content? That is a 00:04:54.939 --> 00:04:57.620 very fair and common concern. But, you know, 00:04:57.839 --> 00:05:00.160 the human review is where the actual value lives. 00:05:00.300 --> 00:05:03.019 Because the AI is just a tool. Right. Claude 00:05:03.019 --> 00:05:05.879 drafts things fast, but you steer the ship. You 00:05:05.879 --> 00:05:08.279 decide what stays and what's deleted. You choose 00:05:08.279 --> 00:05:11.199 the unique angle. You fix the obvious hallucinations. 00:05:11.399 --> 00:05:13.699 Precisely. You add real personal examples to 00:05:13.699 --> 00:05:16.180 the generated text. You make sure the final version 00:05:16.180 --> 00:05:18.639 sounds clear and human. It is a collaboration. 00:05:18.879 --> 00:05:21.139 Yeah. And this is where the repurposing magic 00:05:21.139 --> 00:05:23.839 happens. One authentic YouTube video becomes 00:05:23.839 --> 00:05:26.740 a massive content goldmine. Walk me through that. 00:05:26.740 --> 00:05:28.920 How do we extract that gold mine? Well, you don't 00:05:28.920 --> 00:05:30.939 need a brand new idea every day. You start with 00:05:30.939 --> 00:05:33.899 one real test and reuse the best parts. You feed 00:05:33.899 --> 00:05:36.579 your messy notes back into the AI. Exactly. You 00:05:36.579 --> 00:05:38.480 tell Claude to turn those notes into a short 00:05:38.480 --> 00:05:40.639 script. You instruct it to keep the tone simple. 00:05:40.800 --> 00:05:43.180 Yeah. You ask it to explain the problem and show 00:05:43.180 --> 00:05:45.519 the workflow. You make sure it ends with one 00:05:45.519 --> 00:05:48.480 clear viewer takeaway. Which easily becomes a 00:05:48.480 --> 00:05:50.699 short video clip. Or it could be a LinkedIn post 00:05:50.699 --> 00:05:53.899 or an X thread. How do you avoid burnout if you're 00:05:53.899 --> 00:05:56.040 producing all these different formats? You let 00:05:56.040 --> 00:05:59.300 Claude do the heavy lifting of repurposing. You 00:05:59.300 --> 00:06:02.579 feed your core test back into the system. Claude 00:06:02.579 --> 00:06:04.759 drafts the new formats based on your initial 00:06:04.759 --> 00:06:07.680 work. You let one core experiment automatically 00:06:07.680 --> 00:06:11.300 fuel a dozen smaller content pieces. Exactly. 00:06:11.500 --> 00:06:14.779 It saves so much time. Sponsor. We have the platform 00:06:14.779 --> 00:06:17.699 and required mindset. We know what we are doing 00:06:17.699 --> 00:06:19.920 and where we are doing it. Now we need a reliable 00:06:19.920 --> 00:06:22.959 map. How do we execute this without getting completely 00:06:22.959 --> 00:06:26.410 overwhelmed? Well, the source outlines a very 00:06:26.410 --> 00:06:30.350 specific 30 -day blueprint. It breaks the overwhelming 00:06:30.350 --> 00:06:33.629 task into weekly sprints. Let's look at Week 00:06:33.629 --> 00:06:37.050 1. Week 1 is entirely about finding simple video 00:06:37.050 --> 00:06:39.889 ideas. You aren't building complex automations 00:06:39.889 --> 00:06:42.910 just yet. You give Claude a prompt for 10 easy 00:06:42.910 --> 00:06:45.850 -to -test concepts. Right. These ideas need a 00:06:45.850 --> 00:06:48.009 specific structure, though. A viewer problem, 00:06:48.430 --> 00:06:51.129 an angle, and a clear takeaway. Yes, exactly. 00:06:51.529 --> 00:06:53.850 You review those ten generated ideas, you pick 00:06:53.850 --> 00:06:56.250 the three easiest ones to actually record. Keep 00:06:56.250 --> 00:06:59.569 it simple. Then week two shifts the focus slightly. 00:06:59.949 --> 00:07:02.610 Yeah, week two is about building a highly repeatable 00:07:02.610 --> 00:07:05.189 video format. Why is standardizing the format 00:07:05.189 --> 00:07:07.550 so crucial here? Because it removes friction. 00:07:07.920 --> 00:07:10.139 Your video always starts with a rough idea. You 00:07:10.139 --> 00:07:12.620 show the clawed prompt. You show the raw output. 00:07:12.899 --> 00:07:14.959 You make your edits. You deliver the takeaway. 00:07:15.379 --> 00:07:17.459 Following this logical path makes recording much 00:07:17.459 --> 00:07:19.939 faster. Every single video follows the exact 00:07:19.939 --> 00:07:22.879 same narrative arc. Exactly. Then week three 00:07:22.879 --> 00:07:25.319 is where the dynamic really shifts. You let the 00:07:25.319 --> 00:07:27.680 audience dictate the actual content you make. 00:07:27.920 --> 00:07:29.959 You completely stop guessing what people want 00:07:29.959 --> 00:07:32.740 to see online. Right. You feed viewer comments 00:07:32.740 --> 00:07:35.300 directly into Claude. You take their specific 00:07:35.300 --> 00:07:37.500 questions and their daily struggles. You ask 00:07:37.500 --> 00:07:40.740 the AI to group them into new ideas. It is brilliant. 00:07:41.060 --> 00:07:44.779 Whoa, beat. Imagine scaling from zero to a fully 00:07:44.779 --> 00:07:47.560 automated feedback loop in just four weeks. It 00:07:47.560 --> 00:07:50.360 is an incredibly powerful way to build a business. 00:07:50.720 --> 00:07:53.560 You transition from a creator to a curator. You 00:07:53.560 --> 00:07:55.740 simply build what they are actively asking for. 00:07:55.939 --> 00:07:59.060 Exactly. Then week four is about deep review 00:07:59.060 --> 00:08:02.589 and future planning. You feed your 30 -day performance 00:08:02.589 --> 00:08:05.089 metrics back into Claude. You carefully look 00:08:05.089 --> 00:08:07.689 at your views, your saves, and your comments. 00:08:07.889 --> 00:08:10.430 Saves are a massive signal, by the way. A save 00:08:10.430 --> 00:08:13.050 means your workflow is a real utility. People 00:08:13.050 --> 00:08:15.149 want to return to it later. Yeah. So Claude helps 00:08:15.149 --> 00:08:17.629 you identify which topics hit the hardest. You 00:08:17.629 --> 00:08:19.529 figure out which content ideas you should repeat. 00:08:19.649 --> 00:08:21.769 You learn which concepts could eventually become 00:08:21.769 --> 00:08:24.149 a paid product. Right. What if you get to week 00:08:24.149 --> 00:08:26.910 four and your metrics yield absolutely zero comments? 00:08:27.230 --> 00:08:29.569 Oh, the guide has a great contingency plan for 00:08:29.569 --> 00:08:32.200 that. You use common beginner questions from 00:08:32.200 --> 00:08:35.919 other platforms. You search Reddit or X for AI 00:08:35.919 --> 00:08:38.700 struggles. Exactly, you just answer the questions 00:08:38.700 --> 00:08:40.980 people are asking elsewhere. You simply pivot 00:08:40.980 --> 00:08:43.259 to answering questions people are asking in other 00:08:43.259 --> 00:08:45.759 places. Yeah, you borrow their curiosity until 00:08:45.759 --> 00:08:47.700 you build your own audience. Okay, let's look 00:08:47.700 --> 00:08:50.379 at the horizon here. Let's say after 30 days 00:08:50.379 --> 00:08:52.899 you have an audience. A solid, engaged group. 00:08:53.139 --> 00:08:57.009 They trust your transparent workflow. How do 00:08:57.009 --> 00:08:59.870 we responsibly translate that trust into income? 00:09:00.750 --> 00:09:03.129 We want to avoid ruining the foundation we just 00:09:03.129 --> 00:09:05.549 built. Will you watch closely for buying intent? 00:09:05.820 --> 00:09:08.259 Audience comments will eventually change. They 00:09:08.259 --> 00:09:10.320 stop asking general questions. They start getting 00:09:10.320 --> 00:09:12.980 more specific. Right. They start asking for specific 00:09:12.980 --> 00:09:15.899 assets. They ask if you have a template for your 00:09:15.899 --> 00:09:17.899 workflow. They might ask you to personally help 00:09:17.899 --> 00:09:20.659 them set it up. Exactly. Those comments are actual 00:09:20.659 --> 00:09:23.860 business signals. They want your specific workflow 00:09:23.860 --> 00:09:26.460 to solve their problem. That is when you introduce 00:09:26.460 --> 00:09:29.679 small... digital products. Yes. So you might 00:09:29.679 --> 00:09:32.379 offer a simple template pack first. Maybe price 00:09:32.379 --> 00:09:35.720 it around $19 to $49. People are happily paying 00:09:35.720 --> 00:09:38.700 for clarity and time savings. You can even prompt 00:09:38.700 --> 00:09:41.000 Claude to build this initial pack. You feed it 00:09:41.000 --> 00:09:44.159 your final polished workflow. You ask it to write 00:09:44.159 --> 00:09:46.879 beginner -friendly setup instructions. Exactly. 00:09:46.960 --> 00:09:49.440 It creates the product for your loyal audience. 00:09:49.639 --> 00:09:52.200 We see this mistake all the time, though. Someone 00:09:52.200 --> 00:09:55.259 spends six months building a massive AI course. 00:09:55.480 --> 00:09:58.149 Oh, yeah. It happens constantly. Which is like 00:09:58.149 --> 00:10:00.129 building a whole car before checking if they 00:10:00.129 --> 00:10:02.509 want a steering wheel. Yeah. You are just assuming 00:10:02.509 --> 00:10:05.389 demand. It is a huge trap. Start with the steering 00:10:05.389 --> 00:10:08.309 wheel. The guide also warns against showing only 00:10:08.309 --> 00:10:10.850 polished results. Because people learn much more 00:10:10.850 --> 00:10:13.490 from the messy, realistic process. Exactly. It 00:10:13.490 --> 00:10:15.990 warns against copying random AI trends, too. 00:10:16.289 --> 00:10:18.649 If you haven't tested it, do not share it. Real 00:10:18.649 --> 00:10:20.929 mistakes are far more useful than fake perfection. 00:10:21.149 --> 00:10:24.080 Absolutely. And keep the initial setup very simple. 00:10:24.340 --> 00:10:26.379 You don't need five different AI tools chained 00:10:26.379 --> 00:10:29.299 together. One solid -clawed workflow is plenty 00:10:29.299 --> 00:10:32.100 for a beginner. Making things too complicated 00:10:32.100 --> 00:10:34.860 is a massive pitfall. Yeah, it just scares people 00:10:34.860 --> 00:10:37.659 away. The income growth path feels very organic 00:10:37.659 --> 00:10:40.740 here, though. Simple prompts slowly evolve into 00:10:40.740 --> 00:10:43.740 affordable templates. Then custom setups become 00:10:43.740 --> 00:10:46.740 higher -priced services. A custom setup might 00:10:46.740 --> 00:10:49.980 cost... $100 to $300. Like building a YouTube 00:10:49.980 --> 00:10:52.480 content structure for a fitness creator. Right. 00:10:52.659 --> 00:10:56.019 You prompt Claude for that specific client niche. 00:10:56.399 --> 00:10:59.519 You request 10 video ideas and a posting schedule. 00:11:00.139 --> 00:11:03.460 Eventually, those mature workflows become comprehensive 00:11:03.460 --> 00:11:06.460 paid courses. But it all starts with that first 00:11:06.460 --> 00:11:09.159 messy test. It really does. When is the exact 00:11:09.159 --> 00:11:11.539 right psychological moment to introduce a paid 00:11:11.539 --> 00:11:14.539 offer? When you see repeated audience questions 00:11:14.539 --> 00:11:18.450 asking for the exact same asset. That repetition 00:11:18.450 --> 00:11:21.330 is the undeniable signal of market demand. You 00:11:21.330 --> 00:11:23.269 sell the solution exactly when your audience 00:11:23.269 --> 00:11:25.389 actively asks for it. Yeah, you never have to 00:11:25.389 --> 00:11:27.190 force a sale again. Let's pull back and look 00:11:27.190 --> 00:11:30.509 at the bigger picture. The core philosophy here 00:11:30.509 --> 00:11:32.769 is deeply fascinating. It really changes how 00:11:32.769 --> 00:11:35.289 you view online business. The person who makes 00:11:35.289 --> 00:11:38.179 money in the AI era isn't always building software. 00:11:38.600 --> 00:11:40.740 Often it's the person acting as a relatable guide. 00:11:40.940 --> 00:11:43.340 Because trust is the actual product you are building. 00:11:43.559 --> 00:11:45.480 You are demystifying the technology. You are 00:11:45.480 --> 00:11:47.399 helping people who feel totally left behind. 00:11:47.600 --> 00:11:51.000 You share real, flawed workflows. You watch how 00:11:51.000 --> 00:11:54.519 real viewers respond. Then you slowly turn the 00:11:54.519 --> 00:11:57.500 strongest signals into simple products. It is 00:11:57.500 --> 00:11:59.820 a profound shift in perspective. I have a challenge 00:11:59.820 --> 00:12:03.279 for you this week. Run just one messy, clawed 00:12:03.279 --> 00:12:06.460 workflow test. Document the entire learning process. 00:12:06.500 --> 00:12:08.639 Just put it out there regardless of how unpolished 00:12:08.639 --> 00:12:11.080 it feels. Exactly. It leaves you wondering about 00:12:11.080 --> 00:12:14.690 the future. If AI eventually automates the coding 00:12:14.690 --> 00:12:16.990 of our apps and the writing of our books, will 00:12:16.990 --> 00:12:18.950 the most valuable currency left in the world 00:12:18.950 --> 00:12:21.809 simply be our human ability to curate, stumble, 00:12:22.029 --> 00:12:24.629 and explain the process to each other? Out to 00:12:24.629 --> 00:12:25.090 your own music.