Aren't the first two points mutually contradictory? It seems to me that 'making decisions early' is roughly the same as 'planning'. Why not plunge ahead with a design approach, and give yourself permission to fail?