Knowing What Not to Build

Is what you're building with AI actually what you should be building?

This question, without AI, has always been there. But right now, with AI making everything feel within reach, it matters more than ever.

I'm building a product myself, designing and developing entirely on my own, with AI as my co-pilot. And every single day I wake up with new ideas. Maybe add this. Maybe introduce that. Maybe build this whole other thing that would be kind of cool. AI makes all of it feel achievable. Ship it in a day, maybe two. And that's exactly where it gets dangerous.

Because the question was never "can I build this?" It's always been "should I?" Is this what users actually want? Is this what they need right now? Or is it just a good idea I had at 11pm that feels less exciting by morning? I keep throwing things out. Features I spent hours thinking through. Ideas I was genuinely excited about. Not because I couldn't build them, but because most of them don't belong yet. Or maybe ever.

And this is where real product thinking actually starts. Do users actually want to use your app more? Do they want more features, more screens, more decisions? Probably not. AI should enable us to think further than that. Maybe the move isn't adding features. Maybe it's getting rid of most of the interface altogether. Can we automate the things users shouldn't have to think about? Can we make the system smarter, so people spend less time inside the product and more time getting on with their lives?

Knowing what not to build has always been the skill. AI just makes it more urgent.