I think build as we go, otherwise we over look possible pitfalls. We need trial and error periods to learn as we go and adapt. Waiting leads to a failure to plan, then it might be too late to adapt.