Learning Library

← Back to Library

AI-Powered Micro SaaS Side Hustle

Key Points

  • The current AI era lets anyone turn natural language into functional code, enabling rapid, low‑cost software creation that wasn’t possible just a few years ago.
  • Tools like lovable.dev make it possible to build complete web pages by simply describing what you want, turning software development into a “scalpel” rather than a “hammer.”
  • This shift opens micro‑niche markets—such as custom note‑taking, fantasy football, or event‑planning tools—where a single creator can become the authority and generate sustainable side revenue.
  • Because the technology is ahead of mainstream adoption, early adopters have a limited‑time advantage to capture these untapped niches before the market catches up.
  • The video aims to move beyond vague inspiration by sharing concrete entrepreneurial “war stories” and actionable steps for building an AI‑driven side gig.

Full Transcript

# AI-Powered Micro SaaS Side Hustle **Source:** [https://www.youtube.com/watch?v=8psoB8EFdc8](https://www.youtube.com/watch?v=8psoB8EFdc8) **Duration:** 00:17:50 ## Summary - The current AI era lets anyone turn natural language into functional code, enabling rapid, low‑cost software creation that wasn’t possible just a few years ago. - Tools like lovable.dev make it possible to build complete web pages by simply describing what you want, turning software development into a “scalpel” rather than a “hammer.” - This shift opens micro‑niche markets—such as custom note‑taking, fantasy football, or event‑planning tools—where a single creator can become the authority and generate sustainable side revenue. - Because the technology is ahead of mainstream adoption, early adopters have a limited‑time advantage to capture these untapped niches before the market catches up. - The video aims to move beyond vague inspiration by sharing concrete entrepreneurial “war stories” and actionable steps for building an AI‑driven side gig. ## Sections - [00:00:00](https://www.youtube.com/watch?v=8psoB8EFdc8&t=0s) **Building AI-Powered Side Gigs** - A practical walkthrough showing why the current AI boom lets anyone quickly create niche software side projects—without deep coding skills—by leveraging tools like lovable.dev. - [00:03:28](https://www.youtube.com/watch?v=8psoB8EFdc8&t=208s) **Low‑Code Platform Recommendations** - The speaker explains choosing Lovable.dev for its rapid, continuous feature delivery and recent Stripe integration, and recommends Outa for its all‑in‑one backend services (auth, payments, CRM, email) that seamlessly integrate with Lovable.dev. - [00:07:31](https://www.youtube.com/watch?v=8psoB8EFdc8&t=451s) **Distribution First in AI Ventures** - In the AI era, building a successful side‑gig hinges on prioritizing deep distribution knowledge—understanding where the target community hangs out, their pain points, and earning their trust—before even selecting a product, giving a unique competitive edge over larger model makers. - [00:11:46](https://www.youtube.com/watch?v=8psoB8EFdc8&t=706s) **Finding Non‑AI‑Obsoleted Opportunities** - The speaker advises entrepreneurs to identify enduring niche problems that AI giants won’t replace, leverage micro‑communities for distribution, and build specialized solutions beyond AI’s core consumer stack. - [00:15:12](https://www.youtube.com/watch?v=8psoB8EFdc8&t=912s) **AI Partner for Modern Entrepreneurship** - The speaker encourages entrepreneurs to use AI as a collaborative thinking partner—leveraging prompt guides, micro‑niches, and distribution focus—while emphasizing authentic expertise and fair monetization in today’s uniquely accessible coding era. ## Full Transcript
0:00Okay, this is my video on how to build a 0:03side gig in the age of AI. So many of 0:06the guides out there are generic. 0:08They're not actually practical. We're 0:10going to get super practical and I'm 0:12going to explain to you why this moment 0:14matters and then how you actually go 0:15about building a side gig in the age of 0:17AI. So buckle in. Number one, this is a 0:20unique moment and it won't last forever. 0:23We have an opening and I want to explain 0:25why it works in the market right now. 0:27LMS are making it very very easy to go 0:30from natural language to code. You just 0:32talk and you can get what you want. One 0:34of the tools I'll mention later in this 0:36video is lovable.dev. It's super easy 0:38just to type in what you want and get a 0:40working web page. That is something that 0:42as much as you may already know about it 0:45because you're an AI enthusiast 0:46listening to this video, most people 0:48don't yet. That is going to change. And 0:51during this moment, you have a chance to 0:54build software that would not be 0:57possible to build two years ago, one 1:00year ago. And what I mean by that is 1:02that software has been a hammer instead 1:05of a scalpel for a long, long time. It 1:08cost a lot of money. It cost teams and 1:10teams of developers. You had to go to 1:11Silicon Valley and you had to raise 1:13money to build any kind of software for 1:16most of my career. That is no longer 1:18true. You don't have to do that. You can 1:20build software in nights and weekends 1:22even if you've never coded before. And 1:24that means that you can build custom 1:27software for a specific tiny audience. 1:29If your audience loves to take notes and 1:31they have a particular way of taking 1:33notes and you've always had a 1:34note-taking system and you wanted to 1:36share it with the world, well, you can 1:38do that now. If you wanted to build a 1:40fantasy football software that you have 1:43never been able to find before, you can 1:45do that now. If you wanted to build some 1:47kind of special event planning software 1:49that you would never have been able to 1:50get in the market, well, you can do that 1:52now. You get the idea. These micro 1:54markets exist, but no one has been 1:57building software for them because 1:58software is such a blunt instrument 2:00because it was so expensive. By moving 2:02from natural language to code, we now 2:05have the ability to treat software like 2:08a scalpel. And so we can carve out these 2:10micro niches and build really 2:12sustainable side businesses where we are 2:14the authority on that particular tiny 2:16corner of the universe for that 2:18particular kind of customized software. 2:21It has never been possible before and it 2:23won't last forever. It won't last 2:25forever because this moment is a moment 2:28when the technology is ahead of the 2:29adoption curve. And so you listening to 2:32this video, you're an early adopter. you 2:34have the chance to go out there and look 2:36at a niche that you know well and go 2:38after it. And my goal with this video is 2:40not just to sort of inspire you and give 2:42you generic ideas is to give you war 2:44stories from my experience as an 2:45entrepreneur and also to give you a 2:48sense of how it actually works out there 2:50through my conversation with founders 2:52and others. So if you're starting a 2:54business in 2025, let's assume you get 2:56it, you get the leverage, this is the 2:58moment, etc. What are your tool sets, 3:00right? What do you have to work with 3:02that you didn't have before? Well, I'm 3:04going to name five tools for you, and I 3:07think that you're going to find them 3:08incredibly easy to work with in 3:10combination, and I want to list what 3:12each of them does. This is the absolute 3:14simplest tool set I've been able to come 3:15up with, and I want to explain what each 3:18does and explain why they're in the 3:20stack. Each one earns its keep. Number 3:22one, I mentioned it already, 3:24lovable.dev. 3:25It's no code, it's low code. I want to 3:28give you the straight reason why I 3:29picked this one. this team ships. There 3:32are lots and lots and lots of vibe 3:35coding pieces of software. I could have 3:37recommended Bolt. I could have 3:38recommended Replet. There are lots of 3:41other choices to choose from. I picked 3:43lovable.dev because the team ships fast 3:47and because they are dedicated to making 3:49the product meaningfully better in a in 3:52a cadence of weeks. Like in two weeks, 3:55the product will be better than it is 3:56today. And that's been true for months 3:58and months and months and months. you 3:59can bet on that trajectory. And so, 4:01Lovable.dev offers you a chance to 4:03actually build a functioning site. They 4:05recently launched Stripe integrations. 4:07They have a back-end integration that 4:09works. You can publish to a custom URL. 4:11It is kind of like a web presence in a 4:14box. It's really, really easy. There are 4:16a few other tools though that you may 4:18find useful along the journey. Number 4:20two is uh Outa. Why do I recommend them? 4:23They're not very wellnown actually. I 4:26recommend them because there is one 4:28stack there for everything you need from 4:31a back-end office perspective. So you 4:33can get user authentication that way. 4:35You can get subscription payments that 4:36way. You can get a basic CRM or contact 4:39relationship management database that 4:41way. You can get basic email that way. 4:43It's all under one umbrella which makes 4:45it really really easy if you're starting 4:47out. And it integrates with Love and you 4:49can pull stuff into lovable that way. So 4:51they kind of work together. Now, if 4:53you're getting to a point where you need 4:54to deploy and you don't want to deploy 4:56with Lovable, this is optional, by the 4:58way, because if you're just getting 5:00started, you can kind of just start with 5:02Lovable. But if you want to go a little 5:04farther, you can do continuous builds of 5:07your software system and easy hosting 5:09with a tool called Versel, which is used 5:11all over the world by developers. It's 5:13very, very famous. It's easy to use as 5:15long as you're willing to work with Chat 5:17GPT a little bit on the documentation 5:19side. I want to mention two other tools 5:21here. Again, both of them are optional. 5:23In my view, the only absolutely required 5:25tool if you really want to strip it down 5:27is lovable.dev. And then out is like the 5:30second one if you really want to add 5:31some backend office. And then everything 5:33else is depending on what you want to 5:35build. And if you want to build 5:36something more technical, you add more 5:38technical tooling and so on. Two other 5:40tools to be aware of. Framer offers drag 5:43and drop landing pages and Gemini offers 5:46instant AI analysis for free. That's 5:49right. You can make calls to the Gemini 5:51API for free up to a reasonable weight 5:54rate limit for a tiny business to get AI 5:57analysis and generation. So you can 5:59actually incorporate a free LLM into 6:01your product, which is kind of handy. So 6:04that's my basic tool set. It's super 6:07flexible. It's like a Swiss Army knife. 6:09You can do almost anything with it. And 6:10I want to spend some time talking about 6:14the philosophy of picking a problem now 6:16because in my view everyone is asking me 6:19for the tools. So I gave you the tools. 6:21But the really interesting thing is why 6:24you pick the problem you pick and how 6:26you build on that problem to solve a 6:28customer painoint. This is where sort of 6:30entrepreneur Nate puts on his hat. The 6:33craft of entrepreneurship 6:35actually has changed in important ways 6:37in the age of AI. And so we talk about 6:40the strategic moment. We talk about the 6:41tool set. The tool set that we have as 6:44entrepreneurs, our skill set, the things 6:45that this moment demands from us, that's 6:48also different. First, we need to think 6:50about distribution much much more as 6:53builders than we used to. So before you 6:56would put the product first and you 6:57would make sure the product really 6:58solved the customer problem and then you 7:00would work through established 7:01distribution channels. So if you were in 7:03the consumer space, you would figure out 7:04how to get into stores and so on. Well, 7:06not anymore. Now you have to assume that 7:10someone around you is building something 7:12that is sort of like what you have and 7:14your goal is to get distribution with 7:16the micro niche that you already know 7:18well. And that is why at the top of this 7:20video I recommended that you use a micro 7:23niche that you know well because if 7:25you're already a member of that 7:27community, connecting with them, talking 7:28with them is going to feel natural. 7:30You're going to understand their needs. 7:31You're going to understand their pain 7:32points. You're going to be able to solve 7:33their problems. And you are also going 7:36to have a sense of where they hang out 7:38and how you can reach them. And that's 7:39called distribution. That's distribution 7:41knowhow. And that is an edge. That is an 7:43edge that no major model maker has. And 7:45so part of how you know you can compete 7:48with the likes of open AI and anthropic 7:50is because you have that distribution 7:53knowledge. So the first principle, the 7:55first skill set you need to think about 7:57if you're in the age of AI and you're 7:59building something, you want to build a 8:00side gig, think about your distribution 8:02first. What is your distribution 8:04advantage? What do you know that other 8:06people don't know about your market and 8:09where they hang out and what they like 8:10and what they don't like and what their 8:12pain points are? And are you a member of 8:14that community in such a way that they 8:15trust you and you're respected and you 8:17won't look like you're just scamming 8:18people if you talk about a product that 8:20you built? That is really, really 8:22important. In fact, it's so important 8:24that people building side gigs now pick 8:26the product after they pick the 8:28distribution channel. And that is 8:30completely the reverse of what I was 8:31taught when I got started building and 8:34building companies, building product, 8:36etc. 8:37That's that's really different. Second 8:39thing that's different in the age of AI, 8:41the second principle of entre 8:42entrepreneurship that has changed, we 8:44need to have the skill to know when to 8:49stop building the product. And that is 8:51new because AI will tell you over and 8:53over and over and over again, you can 8:55build more. You can add a database. You 8:57can add a customer relationship 8:58management. You can add a new product. 9:00Expand. Expand. Expand. Expand. You need 9:02to be able to say no. Before, because 9:04software was so expensive, you had to 9:06cut just to sort of find a way to make 9:08it work. Now software is so cheap. It's 9:11very tempting to add an ad and add an 9:12and add. You are the one that has to 9:14develop the skill to say no. You have to 9:17say this is a night and weekend project. 9:18This is the only thing I need to build 9:20to show that I can solve the problem. 9:21That discipline is something that was 9:23imposed by external cost controls 9:25before, not anymore. Now it's on you 9:27now. It's a skill you have to have. Know 9:29when to say no. That is really, really 9:31important. The third major skill that 9:34entrepreneurs bring to the table or must 9:36bring to the table now that they didn't 9:37have to bring before. You need to know 9:40in a very fine grained way where value 9:44lies in your product. You know how I 9:46said software was a blunt hammer before? 9:47Like when you buy Salesforce, Salesforce 9:50might be really crappy in a lot of 9:52different spots, but because it has 9:54existing distribution relationships with 9:56a lot of big companies, the buyer 9:58doesn't really care and so it doesn't 10:00get better. It's different in the micro 10:02niches I'm talking about for side gig 10:04Builders. Your people are not loyal to 10:06you. They will not necessarily buy your 10:08product unless it's good and it really 10:10solves the problem. And so you have to 10:12develop an extraordinary eye for how 10:14monetization and price points and what 10:16people will pay tracks to specific 10:18features in the product. For 10:20entrepreneurs before they could build 10:22the whole product and they might not 10:24know which thing really worked because 10:26the distribution advantage was so blunt 10:29like you you send it off to Target and 10:30like we'll see and like I guess it 10:32works. It sold. Well, now you can tell 10:34where on the page people abandoned and 10:36also you hear from people in Reddit and 10:39Discord chats exactly what they'd like 10:41or dislike about your product and 10:42they're very very specific and you can 10:45see the impact as soon as you change a 10:47feature because suddenly the conversion 10:49rate changes. Now, in theory, that was 10:52all possible before as long as you were 10:54building in software and digital 10:55funnels. But now, it matters more 10:58because people are less loyal and the 11:01market is more fragmented and more 11:03people are building. And so, people are 11:04like very very disloyal when it comes to 11:07customer purchases and what they are 11:09willing to buy and what they're willing 11:10to invest in. Which means you have to be 11:12extremely good at explaining exactly how 11:16your product solves their problem to 11:18earn their trust and loyalty. Now the 11:20good news is if you can do that, if you 11:23can explain why you are passionate about 11:26that micro niche, how your product 11:28really solves a problem they care about, 11:29why the monetization feels fair, like a 11:33square deal to them, you will eventually 11:36earn their trust, and that's where the 11:37distribution advantage starts to become 11:39rock solid in your favor. and it feels 11:42like a tailwind pushing you forward. 11:44You'd love that. You earn that by 11:46building into that distribution wedge 11:48with your micro community and actually 11:51delivering a product that really solves 11:53a problem. And that brings me to the 11:54fourth skill. One of the hardest skills 11:57in the age of AI is finding a problem 12:00that you can be confident AI is not 12:02going to make obsolete. And I want to 12:04spend some time talking about this 12:06because this one is a little bit scary 12:08for people. People often ask, well, 12:10yeah, it's easy to vibe code now, but 12:11why would I build because OpenAI is just 12:14going to take all of the ideas? I don't 12:17think they will. If you think about it, 12:19their strategy very clearly is to drive 12:21a consumer stack. They want you to spend 12:23time as a consumer in OpenAI thinking 12:26and doing your day, but they're not the 12:28best in the world at some of these 12:29peripheral things. Like, they launched a 12:30a meeting noteaker. It's okay. The 12:32meeting note-taker is not going to 12:33compete with the dedicated meeting 12:35notetakers. It never is. That's not 12:37where their focus is. Claude is trying 12:39to eat a lot of the work primitives. 12:40They're trying to eat Excel. They're 12:42trying to eat, you know, claude code. 12:43They're trying to go after PowerPoint. 12:45They want you to spend their day there. 12:47Microsoft should be a little bit worried 12:48about value disintermediation, but as a 12:51as a small builder, you're not that 12:53worried about it. They provide 12:54intelligence for you. That's fine. And 12:55so, you should instead be thinking about 12:59the kinds of pain points that persist 13:02regardless of the intelligence of the 13:04model. What are the things that people 13:06struggle with where a smarter model 13:08wouldn't fix it? So, as an example, if 13:11you are solving a painoint for someone 13:13and it's a coordination problem between 13:15multiple strands of software or multiple 13:17parts of their day or physical and 13:18digital, that is not something that more 13:20intelligence necessarily makes go away. 13:23Another example, if you are solving a 13:25painoint for someone and it bridges the 13:27uh physical world and the digital world. 13:29So you're providing a physical service 13:31mediated digitally or ordered digitally. 13:33That's not necessarily something that 13:34more intelligence makes goes it doesn't 13:37more intelligence doesn't fix that. 13:38Another example, if you are focused on 13:42providing a digital service, but the 13:45digital service is predicated on your 13:48user feeling like they got stuck in a 13:52way that isn't tied to immediate 13:55answers. That is a very powerful place 13:57to solve for. An example of that, let's 14:00say you have to complete a workflow and 14:02you're trying to go through and get all 14:03the steps in your workflow done at work 14:05and maybe it's a a building a product 14:07requirements document. You have to go 14:09through each of those stages and build 14:10it out. You have to get approvals and 14:12then you have to go back and talk to 14:13engineers. That is something that is not 14:15really changing. Now you may have AI 14:18help you do it faster, but the workflow 14:20itself is pretty stable. I want to 14:22suggest to you that one of the things 14:24that is a key for an entrepreneur is to 14:26look at the world in those kinds of 14:27terms. Look at the parts that swap out 14:29when you add more intelligence and look 14:31at the parts that stay steady. In this 14:33case, the workflow stays kind of steady. 14:35You're still going to have senior PMs 14:38wrestling with requirements, wrestling 14:40with engineering on technical 14:41requirements, wrestling with business 14:43stakeholders and trying to figure it out 14:44and trying to make the workflow go 14:45together. I know this because I've lived 14:47it. Well, in that world, the workflow is 14:50a painoint that additional intelligence 14:52doesn't actually solve. In fact, it's a 14:54it's a mega painoint. It yields a bunch 14:55of downstream pain points. Whole 14:57companies are built around that 14:58painoint. find those kinds of pain 15:00points, the workflow pain points, the 15:02physical digital pain points, the pain 15:04points that are not solved by just 15:06adding more intelligence. That requires 15:08some real thinking on your part, right? 15:10You have to you have to use your brain a 15:12little bit and that's okay. You 15:14actually, well, one of the things that's 15:15never been easier is to stretch your 15:17brain because you have a thinking 15:18partner in AI. I'm not suggesting that 15:20you do this by throwing pencils at the 15:22wall the oldfashioned way, the way I 15:24used to. I'm suggesting you use AI as a 15:26partner. And that's why I did a whole 15:27writeup with all of the prompts because 15:30I want you to be able to understand how 15:33to use AI as a thinking partner in this 15:35exercise. And I want you to feel like 15:37the tools that you have, the prompts 15:39that you have can be combined with a 15:42philosophy of entrepreneurship in the AI 15:45age that actually sticks. And so just to 15:48just to recap so you don't forget if 15:50you're starting, one, this is a unique 15:52moment. Code is easier to create than 15:54ever. It won't last forever. Micro 15:56niches are a big part of that. Software 15:57was a blunt hammer before. Three, 16:00please, please remember to go for 16:02distribution. Make sure that you put 16:04distribution at the heart of your 16:06strategy or else you're going to regret 16:08it. Now, when it comes to the skills 16:10that entrepreneurs need to have, you 16:11need to remember, think in terms of 16:14problem spaces that more intelligence 16:16won't solve. You remember, you need to 16:18remember to think in terms of your own 16:20niche expertise and what is authentic to 16:22you. You need to remember to think in 16:24terms of a fair and square deal on 16:26monetization. So you understand the 16:28exact levers that your product offers 16:30and how monetization feels like a square 16:32deal and feels honest and that's how you 16:34build those trusted relationships over 16:36time. You need to think in terms of 16:38bets. Think in terms of a simple MVP 16:42product where you had the discipline to 16:44cut it down to the bones. you can see 16:46that it solves a problem and you're able 16:49to launch it in good time and see if it 16:52works. This micro niche window is not 16:54going to last forever. Your next weekend 16:56project can get something out the door 16:58that actually earns you money. And I'm 17:01not going to pretend that your next 17:02weekend project is going to earn you a 17:04billion dollars, but I think it's fair 17:06if you actually discipline yourself and 17:08start building on nights and weekends to 17:10get to a point where you can build a 17:12little piece of software that gets you 17:13into the hundreds and thousands of 17:14dollars a month. That is totally doable. 17:16I know lots of people who have done 17:17that. I know people who have scaled past 17:19that into five figures a month. You can 17:21do it. You just have to make sure that 17:23you understand how entrepreneurship 17:25actually works today, what the 17:27principles are, where the opportunities 17:28are, and what the toolkit actually looks 17:30like. I hope that this has been helpful. 17:32The full write up with all the links is 17:35on the Substack. And yeah, I hope you 17:38have a good time. I love building. It 17:40has never been a better time to build. 17:42And good luck out there. Good luck 17:44building. I hope you come back and let 17:46me know what you are working on.