I constantly get questions from developers about what they should build and where the opportunities are.
I’ve been thinking about this a lot over the past two weeks.
Building software has never been easier, especially with AI. But the fundamentals of building something people actually want to use and pay for haven’t changed.
You need to really understand the problem you’re solving. Not just have an idea, but know the problem better than most people. That’s how you build a product that just makes sense to your users because you understand what they need and how they think.
If you’re building a product around data, having access to information that isn’t publicly available to AI agents can be a huge advantage. The harder that data is to get, the more valuable your product can become.
And please, don’t just ship the default Claude-generated design. It makes your product look like everything else being built with AI and can cheapen its perceived value before someone even tries it. Design matters. Taking the time to create a custom experience forces you to think about your customers, how they behave, and what makes your product feel natural to them. Good design is part of understanding your users, not just making things look pretty.
I’d also encourage developers to find their first customer before writing a single line of code. Try to land a pilot customer who can be a design partner from day one. Someone who will actually use what you’re building, give you honest feedback, and help you figure out what matters most before launch.
And probably the biggest thing I’ve learned is you don’t need to build a product that does everything.
Just be great at one thing.
It’s perfectly fine if that’s all your product does for the first few years. You’d be surprised how many people will pay for a product that solves one problem extremely well.
A product that does one thing really well can be a much bigger business than one that does ten things okay.