My feed is full of weekend SaaS launches lately.
Someone builds an AI tool in 48 hours, ships a landing page, posts a screenshot of a clean dashboard, and calls it a startup. The stack is usually some combination of Next.js, Supabase, Vercel, and an AI coding assistant.
And to be fair, some of these projects are genuinely impressive.
But I've started paying less attention to how quickly something was built and more attention to what happens after it goes live.
Because here's the uncomfortable part: building a working product is becoming a lot less impressive when almost everyone can do it.
That doesn't mean software engineering is suddenly worthless. It means shipping an MVP isn't the same competitive advantage it used to be.
A polished demo isn't product-market fit
Imagine spending a weekend building an AI-powered PDF tool.
It uploads documents, summarizes content, translates text, and has a surprisingly decent dashboard. You deploy it, share it on X and Reddit, maybe submit it to Product Hunt, and start watching the analytics.
A few dozen visitors. Some signups. A couple of nice comments.
Zero paying customers.
So you do what developers naturally do: add more features.
Maybe the tool needs better exports. Or a different model. Or a nicer interface. Another weekend goes into development, and suddenly you've got twice as many features without any clearer idea of whether people want the product.
This is the part AI doesn't really solve.
It can make adding features ridiculously cheap. It can't tell you whether those features matter to anyone.
And when development gets cheaper, it's surprisingly easy to confuse activity with progress.
The boring stuff starts after launch
Developer communities are great at discussing stacks. Which framework is faster? Should you use Postgres or something simpler? Is Cursor better than Claude Code? Can you get away with a free hosting tier?
I enjoy those conversations too.
But once a product has actual users, the problems change.
Someone can't log in. An email never arrives. Your latest deployment breaks an existing feature. A free user runs through more AI tokens than expected. A paying customer wants a refund.
None of those problems makes for an exciting launch thread.
If you're building alone, though, they're all your problems.
And this is where the idea of a one-person business gets interesting. AI may let you work without a large team, but it doesn't eliminate the work that team would normally do.
You're still the developer, product manager, support person, and whoever gets to figure out why the monthly costs are higher than expected.
The bottleneck simply moves.
I'd rather see ten paying users than another feature list
One thing I think indie developers should do earlier is test whether people are actually willing to pay.
Not just sign up. Not just leave a nice comment. Actually pay.
A simple product that earns $10 a month from ten customers tells you something that a beautifully designed dashboard with zero revenue doesn't.
Of course, getting that first payment means setting up a way to collect money. Depending on the product and where your customers are, that might involve cards, a Merchant of Record, local payment methods, or stablecoins.
If your users already hold USDT or USDC, for example, offering stablecoin payments might be useful.
I'm involved with JangoPay, a project building stablecoin payment tools for online businesses. We work on things like Payment Links and payment integrations, so I've spent a fair amount of time thinking about the payment side of small online products.
But here's the thing: I wouldn't tell a new founder to spend weeks optimizing checkout before figuring out whether anyone wants what they're selling.
Payment infrastructure can help you collect revenue. It cannot create demand.
That's a much harder problem.
A one-person company is still a real company
I like the idea of solo founders. Building a sustainable business without needing to raise money or manage a large team is genuinely appealing.
But I think the current excitement around AI sometimes makes it sound easier than it is.
AI can write code, but you're still responsible when that code breaks. It can generate landing page copy, but someone still has to figure out where customers come from. It can help draft support replies, but it can't make every unhappy customer disappear.
And as building software becomes easier, competition gets tougher.
If five people can build similar products in a week, your advantage probably isn't going to be that you wrote the code faster. It might be that you understand a particular group of customers better, that your product is more reliable, or that people trust you enough to keep paying every month.
None of that comes automatically from a good prompt.
Shipping is the beginning, not the finish line
I still think AI-assisted development is one of the most exciting things happening in software. It gives people who wouldn't otherwise have the time or resources a much better chance of trying an idea.
That's a big deal.
But the part I'm increasingly interested in isn't the weekend build. It's the months afterward.
Where did the first customers come from? What made them stick around? How did the founder decide what to charge? How much time goes into support? Is the product earning more than it costs to operate?
Those are the stories I'd like to see more of.
Because a successful build and a successful business are not the same thing.
A working app is a technical achievement. A product people keep paying for is a business.
If you're working on a side project, how do you decide whether it's worth continuing? Signups, retention, or the first real paying customer?
