Agreed. You realize that making money is the hard part so there's no need to do a ton of thinking into technical decisions until it becomes "a good problem" to have.

These days I might have an idea for a feature in my head and pitch it to people as if it exists to gauge interest. If it seems like it's a clear winner I build it into my product. If not, I'm not writing any code.