Nate B Jones explains how to build an AI-powered side gig using no-code tools and entrepreneurial strategy
Solo presentation by Nate B Jones on building a profitable side business using AI tools in 2025.
Summary
Nate B Jones argues that the current moment — where AI makes it possible to go from natural language to working software — represents a narrow, time-limited window for entrepreneurs to build micro-niche software businesses that were previously impossible. He walks through a practical five-tool stack centered on Lovable.dev and outlines four key entrepreneurial skills that have fundamentally changed in the AI age. His central argument is that distribution knowledge, not technical skill, is now the primary competitive advantage for side-gig builders, and that finding problems AI intelligence alone cannot solve is the critical strategic challenge.
Key Takeaways
FULL TRANSCRIPT
Why This Moment Is Unique for Side-Gig Builders
Nate B Jones: This is my video on how to build a side gig in the age of AI. So many of the guides out there are generic — they're not actually practical. We're going to get super practical. I'm going to explain to you why this moment matters and then how you actually go about building a side gig in the age of AI.
Number one: this is a unique moment and it won't last forever. We have an opening, and I want to explain why it works in the market right now. LLMs are making it very easy to go from natural language to code. You just talk and you can get what you want. One of the tools I'll mention later in this video is Lovable.dev. It's super easy to just type in what you want and get a working web page. That is something that — as much as you may already know about it because you're an AI enthusiast listening to this video — most people don't yet. That is going to change.
During this moment, you have a chance to build software that would not have been possible to build two years ago, one year ago. What I mean by that is that software has been a hammer instead of a scalpel for a long, long time. It cost a lot of money. It required teams and teams of developers. You had to go to Silicon Valley and raise money to build any kind of software for most of my career. That is no longer true. You don't have to do that. You can build software on nights and weekends even if you've never coded before.
That means you can build custom software for a specific, tiny audience. If your audience loves to take notes and they have a particular way of taking notes and you've always had a note-taking system you wanted to share with the world, you can do that now. If you wanted to build fantasy football software that you've never been able to find before, you can do that now. If you wanted to build some kind of special event planning software that you would never have been able to get in the market, you can do that now. You get the idea.
These micro-markets exist, but no one has been building software for them because software has been such a blunt instrument — because it was so expensive. By moving from natural language to code, we now have the ability to treat software like a scalpel. We can carve out these micro-niches and build really sustainable side businesses where we are the authority on that particular tiny corner of the universe for that particular kind of customized software. It has never been possible before, and it won't last forever.
It won't last forever because this moment is a moment when the technology is ahead of the adoption curve. You, listening to this video, are an early adopter. You have the chance to go out there, look at a niche that you know well, and go after it. My goal with this video is not just to inspire you and give you generic ideas — it's to give you war stories from my experience as an entrepreneur and also to give you a sense of how it actually works out there through my conversations with founders and others.
The Five-Tool Stack
So if you're starting a business in 2025, let's assume you get it — you get the leverage, this is the moment, and so on. What are your tool sets? What do you have to work with that you didn't have before? I'm going to name five tools, and I think you're going to find them incredibly easy to work with in combination. This is the absolute simplest tool set I've been able to come up with, and I want to explain what each does and why they're in the stack. Each one earns its keep.
Number one — I mentioned it already — is Lovable.dev. It's no-code, it's low-code. I want to give you the straight reason why I picked this one. This team ships. There are lots and lots of vibe-coding pieces of software. I could have recommended Bolt. I could have recommended Replit. There are lots of other choices. I picked Lovable.dev because the team ships fast and because they are dedicated to making the product meaningfully better on a cadence of weeks. In two weeks, the product will be better than it is today. That's been true for months and months. You can bet on that trajectory.
Lovable.dev offers you a chance to actually build a functioning site. They recently launched Stripe integrations. They have a back-end integration that works. You can publish to a custom URL. It is kind of like a web presence in a box. It's really, really easy.
There are a few other tools that you may find useful along the journey. Number two is Outa. Why do I recommend them? They're not very well known, actually. I recommend them because there is one stack there for everything you need from a back-end office perspective. You can get user authentication that way. You can get subscription payments that way. You can get a basic CRM — contact relationship management — database that way. You can get basic email that way. It's all under one umbrella, which makes it really easy if you're starting out. And it integrates with Lovable, so you can pull things in that way. They kind of work together.
Now, if you're getting to a point where you need to deploy and you don't want to deploy with Lovable — this is optional, by the way, because if you're just getting started you can start with Lovable — but if you want to go a little further, you can do continuous builds of your software system and easy hosting with a tool called Vercel, which is used all over the world by developers. It's very famous. It's easy to use as long as you're willing to work with ChatGPT a little bit on the documentation side.
I want to mention two other tools. Both of them are optional. In my view, the only absolutely required tool if you really want to strip it down is Lovable.dev. Outa is like the second one if you really want to add some back-end office. Everything else depends on what you want to build. If you want to build something more technical, you add more technical tooling and so on.
Two other tools to be aware of: Framer offers drag-and-drop landing pages, and Gemini offers instant AI analysis for free. You can make calls to the Gemini API for free up to a reasonable rate limit for a tiny business to get AI analysis and generation. So you can actually incorporate a free LLM into your product, which is quite handy.
That's my basic tool set. It's super flexible — like a Swiss Army knife. You can do almost anything with it.
The Philosophy of Picking a Problem
I want to spend some time talking about the philosophy of picking a problem, because in my view everyone is asking me for the tools. So I gave you the tools. But the really interesting thing is why you pick the problem you pick and how you build on that problem to solve a customer pain point.
This is where entrepreneur Nate puts on his hat. The craft of entrepreneurship has actually changed in important ways in the age of AI. We've talked about the strategic moment and the tool set. The skill set — the things that this moment demands from us — that's also different.
Skill One: Distribution First
First, we need to think about distribution much more as builders than we used to. Before, you would put the product first, make sure the product really solved the customer problem, and then work through established distribution channels. If you were in the consumer space, you'd figure out how to get into stores and so on. Not anymore.
Now you have to assume that someone around you is building something similar to what you have, and your goal is to get distribution with the micro-niche that you already know well. That is why at the top of this video I recommended that you use a micro-niche that you know well — because if you're already a member of that community, connecting with them and talking with them is going to feel natural. You're going to understand their needs, their pain points. You're going to be able to solve their problems. And you're going to have a sense of where they hang out and how you can reach them. That's called distribution knowhow. And that is an edge.
That is an edge that no major model maker has. Part of how you know you can compete with the likes of OpenAI and Anthropic is because you have that distribution knowledge. So the first principle — the first skill set you need to think about if you're building a side gig in the age of AI — is to think about your distribution first. What is your distribution advantage? What do you know that other people don't know about your market, where they hang out, what they like and don't like, what their pain points are? And are you a member of that community in such a way that they trust you, you're respected, and you won't look like you're just scamming people if you talk about a product you built? That is really, really important.
In fact, it's so important that people building side gigs now pick the product after they pick the distribution channel. That is completely the reverse of what I was taught when I got started building companies. That's really different.
Skill Two: Knowing When to Stop Building
The second thing that's different in the age of AI — the second principle of entrepreneurship that has changed — is that we need to have the skill to know when to stop building the product. That is new, because AI will tell you over and over again: you can build more. You can add a database. You can add a CRM. You can add a new product. Expand, expand, expand. You need to be able to say no.
Before, because software was so expensive, you had to cut just to find a way to make it work. Now software is so cheap, it's very tempting to keep adding. You are the one who has to develop the skill to say no. You have to say: this is a nights-and-weekends project. This is the only thing I need to build to show that I can solve the problem. That discipline was imposed by external cost controls before. Not anymore. Now it's on you. It's a skill you have to have. Know when to say no.
Skill Three: Understanding Where Value Lives
The third major skill that entrepreneurs must bring to the table now — that they didn't have to bring before — is knowing in a very fine-grained way where value lies in your product. Remember how I said software was a blunt hammer before? When you buy Salesforce, Salesforce might be really lacking in a lot of different spots, but because it has existing distribution relationships with a lot of big companies, the buyer doesn't really care and so it doesn't get better.
It's different in the micro-niches I'm talking about for side-gig builders. Your people are not loyal to you. They will not necessarily buy your product unless it's good and it really solves the problem. So you have to develop an extraordinary eye for how monetization, price points, and what people will pay tracks to specific features in the product.
Before, entrepreneurs could build the whole product and might not know which thing really worked because the distribution advantage was so blunt — you send it off to Target and see what happens. Now you can tell where on the page people abandoned. You hear from people in Reddit and Discord chats exactly what they like or dislike about your product, and they're very specific. You can see the impact as soon as you change a feature because suddenly the conversion rate changes.
Now, in theory, that was all possible before as long as you were building in software and digital funnels. But now it matters more because people are less loyal, the market is more fragmented, and more people are building. People are very disloyal when it comes to customer purchases and what they're willing to invest in. Which means you have to be extremely good at explaining exactly how your product solves their problem to earn their trust and loyalty.
The good news is that if you can do that — if you can explain why you are passionate about that micro-niche, how your product really solves a problem they care about, why the monetization feels like a square deal — you will eventually earn their trust. That's where the distribution advantage starts to become rock solid in your favor, and it feels like a tailwind pushing you forward. You earn that by building into that distribution wedge with your micro community and actually delivering a product that really solves a problem.
Skill Four: Finding Problems AI Won't Make Obsolete
That brings me to the fourth skill. One of the hardest skills in the age of AI is finding a problem that you can be confident AI is not going to make obsolete. I want to spend some time talking about this because it's a little bit scary for people. People often ask: yeah, it's easy to vibe-code now, but why would I build when OpenAI is just going to take all the ideas?
I don't think they will. If you think about it, their strategy very clearly is to drive a consumer stack. They want you to spend time as a consumer in OpenAI thinking and doing your day, but they're not the best in the world at some of these peripheral things. They launched a meeting note-taker — it's okay, but it's not going to compete with the dedicated meeting note-takers. That's not where their focus is. Claude is trying to eat a lot of the work primitives — trying to eat Excel, going after PowerPoint. They want you to spend your day there. Microsoft should be a little worried about value disintermediation, but as a small builder, you're not that worried about it. They provide intelligence for you. That's fine.
So you should instead be thinking about the kinds of pain points that persist regardless of the intelligence of the model. What are the things that people struggle with where a smarter model wouldn't fix it?
As an example: if you are solving a pain point for someone and it's a coordination problem between multiple strands of software, or multiple parts of their day, or between the physical and digital worlds — that is not something that more intelligence necessarily makes go away. Another example: if you are solving a pain point that bridges the physical world and the digital world — you're providing a physical service that is mediated or ordered digitally — more intelligence doesn't fix that. Another example: if you are focused on providing a digital service, but the digital service is predicated on your user feeling stuck in a way that isn't tied to immediate answers. That is a very powerful place to solve for.
An example of that: let's say you have to complete a workflow and you're trying to get all the steps done at work — maybe building a product requirements document. You have to go through each of those stages and build it out, get approvals, go back and talk to engineers. That is something that is not really changing. You may have AI help you do it faster, but the workflow itself is pretty stable.
One of the things that is key for an entrepreneur is to look at the world in those terms. Look at the parts that swap out when you add more intelligence, and look at the parts that stay steady. In this case, the workflow stays kind of steady. You're still going to have senior PMs wrestling with requirements, wrestling with engineering on technical requirements, wrestling with business stakeholders, trying to make the workflow come together. I know this because I've lived it. In that world, the workflow is a pain point that additional intelligence doesn't actually solve. In fact, it's a mega pain point — it yields a bunch of downstream pain points. Whole companies are built around that pain point.
Find those kinds of pain points: the workflow pain points, the physical-digital pain points, the pain points that are not solved by just adding more intelligence.
That requires some real thinking on your part. You have to use your brain a little bit — and that's okay. One of the things that's never been easier is to stretch your brain because you have a thinking partner in AI. I'm not suggesting you do this by throwing pencils at the wall the old-fashioned way, the way I used to. I'm suggesting you use AI as a partner.
Recap and Final Thoughts
Just to recap so you don't forget: one, this is a unique moment — code is easier to create than ever, it won't last forever, and micro-niches are a big part of that. Software was a blunt hammer before. Two, please remember to go for distribution. Make sure you put distribution at the heart of your strategy or you're going to regret it.
When it comes to the skills entrepreneurs need to have in the age of AI: think in terms of problem spaces that more intelligence won't solve. Think in terms of your own niche expertise and what is authentic to you. Think in terms of a fair and square deal on monetization — understand the exact levers your product offers and how monetization feels honest, because that's how you build trusted relationships over time. Think in terms of bets. Think in terms of a simple MVP product where you had the discipline to cut it down to the bones, so you can see that it solves a problem and you're able to launch it in good time and see if it works.
This micro-niche window is not going to last forever. Your next weekend project can get something out the door that actually earns you money. I'm not going to pretend that your next weekend project is going to earn you a billion dollars, but I think it's fair — if you actually discipline yourself and start building on nights and weekends — to get to a point where you can build a little piece of software that gets you into the hundreds and thousands of dollars a month. That is totally doable. I know lots of people who have done that. I know people who have scaled past that into five figures a month.
You can do it. You just have to make sure that you understand how entrepreneurship actually works today, what the principles are, where the opportunities are, and what the toolkit actually looks like. It has never been a better time to build.