Without the fantasy, the reality would never come.
Dopamine can be motivating, but over a threshold it's satisfying, which is essentially the opposite of motivating.
I think what's really being warned about is that if you dream too much it becomes satisfying. If you allow yourself to experience the pleasure of fantasising about your future, then you lose the motivation to actually realise that future. Instead of training your neural pathways to more effectively improve your circumstances, you train your neural pathways to get to that satisfaction quicker by simply imagining and believing you will have done so in the future.
The article is discussing fantasizing that does not result in action
I actually find it useful to, instead of immediately acting, sometimes take a bit of time to ask: how do I want my UI to look? How are people going to use this software? What are some typical use case scenarios? I've gained valuable insight by daydreaming a little bit, and not acting immediately.
IMO, you need a mixture of both some planning and also actively working on the project. You can get stuck in fantasies, design something overly complex that you think will solve all use cases (but actually does many things poorly instead of doing one thing well). You can also fantasize about a project so much that you get bored of the project before you even really get started.