This depends on whether you are dealing with processes subject to entropic decay (they break apart and "die" without effort-input) or entropic growth (they optimize under their own power). For the former case, the Nirvana fallacy remains a fallacy; for the latter case, you are in deep trouble if you try to go with the first "good enough" alternative rather than defining a unique best solution and then trying to hit it as closely as possible.