Seeing this question made me think of an interesting one: all the answers are based on the assumption that there are users. But what if there are no users at al

Seeing this question made me think of an interesting one: all the answers are based on the assumption that there are users. But what if there are no users at all — how should tasks be prioritized? Should you continue developing new features or improving existing ones to attract users, or should you start a new project instead?