Changing your idea frequently is perfectly fine and likely healthy (as long as you are not going around in circles). It suggests that you are responding to the competitive landscape, adjusting to use cases, testing hypothesis and tuning what you might build based on what is feasible (given technical, skill, and financial constraints).
Changing your product frequently is, more often than not, very unhealthy. It increases the likelihood that your code will be messy, the user experience confusing and your efforts used inefficiently.
So, if your idea of what to do is changing once a week, the right time to start building the product is when
1. Your idea matures. You will know you are ready when the idea remains mostly unchanged after a couple of weeks.
2. You identify a core feature that is common to many of your ideas over the last few weeks. Often similar set of features can support quite different ideas. Maybe you have been “thinking too big”; making the idea too complex (and thus unstable).* Think in terms of the Minimum Viable Product (MVP)