How to get AI to help with your side hustle

AI coding tools have changed who gets to build: here’s how to turn your domain knowledge into a small product people actually use

Jun 28, 2026

I read a fascinating post by the team at Stripe, titled The age of the solopreneur

This chart shows that more businesses are reaching $1M payment volume faster than ever before, run primarily by solo founders using AI.

Stripe Economics, “The age of the solopreneur”, June 2026. Source: Stripe internal data for businesses with at least $1K in cumulative payment volume.

Here is the key quote:

…the evidence points toward a structural increase in genuine small business activity over 2025 and 2026, driven by solopreneurs. Preliminary evidence of growth in AI tool usage and solopreneur growth in high-AI-adoption sectors suggests that advances in AI are responsible for a meaningful portion of this growth.

Read the full article. It’s fascinating, and the structural shift aligns with what I’m seeing anecdotally, more people are interested in side hustles as well.

Maybe you have an idea for a new finance app?
Maybe you wanted to build a new marketplace?
Maybe you want to earn a bit of passive income.

AI has really opened up the possibilities for all of the above.

Why Now, and Why You

Let me explain why this moment is different from every other “anyone can build an app” moment you’ve heard about over the last decade.

Those previous waves:

no-code tools
website builders
drag-and-drop app makers

They were real, but they had a ceiling. You could build something that looked like a product, but the moment you needed it to do something slightly custom, you hit a wall. You needed a developer again.

What changed is AI coding tools.

Cursor
Claude Code
Codex
Replit

These tools let you describe what you want in plain English and they write the code. When something breaks, you describe the problem and they fix it. You’re learning to describe what you want clearly and to iterate on the result. The good news is that this is a skill you already have, you’ve been doing this your entire career with designers, agencies, contractors, junior staff.

Most people typically know how to look at something and say “that isn’t right, here’s what I actually meant.”

The second reason this moment is different is you. Specifically, the fact that you’re a person with deep domain knowledge, a network, and the ability to make decisions quickly. Those things matter enormously when you’re building a product. You know what problems are real because you’ve lived them. You know who to call to validate an idea because you’ve been in the industry for five or fifteen years. And you don’t need to ask permission from anyone to spend a weekend on something. That combination is genuinely powerful.

Ground Rules Before You Start

Try to solve a problem you’ve personally experienced. It doesn’t have to be in your industry, but you should have felt the pain yourself.

Start with this book, you can read it on Amazon here.

That phrase “fall in love with the problem, not the solution” is really important when you are transitioning from your day job to a side hustle or project. You will probably be solution focused. Falling in love with the problem is harder than it sounds, as noticing problems is harder than it sounds.

You’ll probably build a few things before you find the one that sticks

There are ideas that I think are awesome and then I try them and they kind of suck. So you usually have to build more ideas than you think.

If you want things to practice on, then work on little problems that you experience every day. Maybe that is building a better gym training app and designing out all the stuff that annoys you about it. Maybe it’s making a protein counter because you are trying to get muss as. Maybe it’s something that takes your Strava routes and suggests you better ones with less hills.

A simple one I started with is an easy way to make blog images. It’s a tiny simple tool and you paste in a book and it formats the picture nicely for a blog post. www.blogwritingtoolkit.com

This was fun and I solved a simple problem and I got a result pretty fast. Most people make a mistake I think of trying to make their first thing way too massive.

Give it six months before you decide if you are enjoying it

The first month is novelty, everything is amazing, you can’t believe this works.
The second month is frustration, things break, you can’t figure out why, you start to wonder if this was a stupid idea.
Months three through six are where you actually find out if this is for you. If you’re still thinking about it in the shower at month four, you’ve got something.

Most ideas for me never pass the 6 month test. One that has passed the test has been writing and anything to do with books. I love books, so it doesn’t feel like work. I believe you have to solve your own problems that you actually enjoy to get past the 6 month test.

Don’t think about money yet (with an asterisk*)

This one is interesting, if you are solving a problem in your industry and you know for sure you can make some cash from it, then think about money. That is your unique advantage.

If it falls outside of that, then you are better off getting something that people love first and to solve a real pain point first, before you wonder about making money,

I know that feels counterintuitive. You’re an executive, you think in business cases. But money comes from solving a real problem for real people, and if you start with “how do I monetise this,” you’ll optimise for revenue before you’ve figured out if anyone cares.

Solve the problem first. The money question answers itself later. It’s way harder to do this and to make money than you think.

Do it because it’s fun!

I think that this part is underrated

You already have a job. This should be the thing you’re excited to work on at 9pm on a Tuesday. If it feels like a second job, you’ve picked the wrong problem.

I think the whole market is saturated with people just wondering about how to get users or money and forgetting to do what they actually enjoy.

That’s why side project is actually a better word than side hustle.

Finding Your Idea

Read this essay by Paul Graham.

The way to get startup ideas is not to try to think of startup ideas. It’s to look for problems, preferably problems you have yourself.
The very best startup ideas tend to have three things in common: they’re something the founders themselves want, that they themselves can build, and that few others realize are worth doing. Microsoft, Apple, Yahoo, Google, and Facebook all began this way.

Get ChatGPT to interview you as though it is Paul Graham.

A lot of people get stuck here. They know they want to build something but they can’t land on what.

This part is the hardest. AI makes it a lot easier.

There’s also a great startup idea validation agent you can run: github.com/MaxKmet/idea-validation-agents.

Descope Until It Hurts

This is the most important section in this essay. Read it twice (don’t worry I need to read this section twice as well, it’s hard to descope!)

Your biggest problem will be building too much.

AI coding tools make it so easy to add features that you’ll add twelve things in your first weekend.

A dashboard.
User accounts.
Email notifications.
An admin panel.
Analytics.
A settings page.
And suddenly your simple idea is a bloated mess that does everything and nothing well.

The discipline is in cutting. What is the one thing your product does? The single action that makes it useful? Everything else is a distraction.

Think how simple ChatGPT is, it’s just a chatbox. It’s so clever. Same with Google. Often so many great products are actually extremely simple.

Here’s a trick that you can do.

Tell your AI coding tool to act as Y Combinator during office hours. Tell it your idea and ask it to be brutally honest about what you should cut. Ask it to keep cutting until there’s only one core feature left. That feature is your MVP.

Picking Your Tool

You need a coding tool. Here’s what I recommend.

Cursor is the best tool for this work. Composer 2.5 (an AI model in Cursor) is genuinely impressive. You open it, you describe what you want in plain English, it builds it, you tell it what to change. The learning curve is steeper than the others, but you’ll outgrow the simpler tools quickly, so you might as well start here.

If Cursor feels like too much on day one, start with Codex. It’s included in your ChatGPT subscription, it takes plain English and turns it into working code, and it’s probably the gentlest way in. You can always move to Cursor later.

I wrote a whole blog post about Codex.

Replit is good for getting something running in an afternoon. The trade-off is that it gets expensive as your project grows and is more limited than the alternatives. Think of it as training wheels, useful for your first project, something you’ll graduate from.

Your First Conversation With Cursor/Codex/Replit

Here’s what you type. Something like this:

“I want to build a web app that does [your one thing]. It should have a simple clean interface. The user should be able to [the one action your MVP does]. Use Next.js and React. Keep it simple.”

That’s it. Cursor will generate a project structure, write the code, and you’ll have something you can see in your browser. It won’t be pretty. It probably won’t be exactly right. And that’s fine, because now you have something to react to.

And reacting to something is what you’re good at. You look at it and say “the button should be bigger” or “I want this to show a list instead” or “can you add a field for the user’s email.” Each time, Cursor rewrites the code. You’re having a conversation, and the product evolves through that conversation.

Shipping It: Getting Your Product on the Internet

I reckon it’s a really fun feeling getting your project on the internet and buying a website. I like to do that first!

To do that you need two accounts.

GitHub is where your code lives. Go to github.com and create an account.

Vercel is what puts your site on the internet. Go to vercel.com and create an account. Connect it to your GitHub account. This takes about two minutes.

Then you make a project in Vercel.

Once you have done that go to https://vercel.com/domains and you can buy a domain for about $10 USD a year. Don’t buy domains that cost more than that when you are starting.

It’s also fun having a website, you can check it on your phone and even if you just build something small to start, it’s really cool!

How Much to Budget

$60 USD a month on Cursor with Composer 2.5 will get you a long way. Realistically, once your site has a lot of features and you’re iterating fast, expect $100 to $200 a month. GitHub, Vercel, and Neon are all free at the scale you’ll be operating at.

So your total cost to run a real product on the internet is somewhere between $60 and $200 a month. Think about what that number would have been three years ago.

The Feedback Loop

Once you’ve got something working and it’s live on the internet, give it to three people. Three.

Don’t post it on LinkedIn. Don’t ask for feedback in a group chat. Pick three people who have the problem you’re solving and ask them to try it.

Don’t ask “what do you think?” That question gets you politeness. Ask “can you try to do [the thing]?” and then watch what happens. Where they hesitate, where they get confused, where they tap the wrong button, where they give up, that’s all great information. That’s your roadmap for what to build next.

After three people, you’ll have a clear picture of what’s working and what isn’t. Fix the things that confused people. Then give it to three more. This is the loop. Build, watch, fix, repeat.

You’d be amazed how even with 3 people, the main thing you will encounter is apathy. Most people don’t care.

If you want someone to give you honest feedback, add me on LinkedIn or email me chris@cubdigital.co.nz - (honest feedback is usually harder to get than you think!)

How to know if your product is actually good

If someone literally goes I want to pay you $100, please let me pay you, then that is good.

If someone invites their friends without you asking and then their friends invite their other friends without you asking then that is good.

These are all good signs.

Most likely this won’t happen, you will just be met with, “oh that’s kind of cool.”

The Real Cost

At the end of the day, given that AI has made it so much cheaper than it used to be, the real cost is your time.

“Startups work because founders make them work” is a quote that I love by Sam Altman.

Most likely your first thing will suck more than you think, people will hate it more than you think and then you have a choice. Do you want to make this work or not? Sometimes it’s that the problem isn’t that interesting or again people are just too busy.

That’s why you have to enjoy it.

That’s why it should be fun and solve your own problem. Falling in love with the problem is harder than you think it is!

I spent my first 8 months of writing getting 20 views a month. I didn’t care, I really enjoyed it. 2 years in and I still love it as much as I did on day 1. I haven’t missed a single week in 2 years. I couldn’t even tell you how long I spent writing because I don’t even notice.

The fun upside

Even if no one ever uses your product and everyone hates it, don’t worry there is still heaps of upsides to doing a side hustle/project.

Creativity - it’s so fun! You get to make things and now you don’t need to raise thousands of dollars and convince investors or talk to someone technical to do so.
It helps you to find your passion.
You get to meet new people that share your passion.
You start noticing the world differently, through the lens of a problem solver.
It will probably help you in your job. I’ve had 4 or 5 times where I’ve built a random as side project and then used it for a client 6 months later.