This mantra no longer makes sense. In fact, I'd say that the key to success is to hold off on writing code as long as you can. There are, of course, exceptions to this key, but those exceptions are mostly about technical problems. When it comes to business problems, problems that involve users and workflows, it's best to exercise discretion on the coding part.
Way too many projects get started before the team knows what problem they are trying to solve. We don't understand the user, the workflow, the contingencies, the inertia, the environment, and the true motivations.
It feels good to start coding. It feels like progress, like something meaningful and real is happening. The problem, however, is that real can become real expensive in a real hurry.
So the mantra now is "Stay in Design Mode until you can't take it anymore."