Although there is fierce debate about how fast natural selection can proceed and if natural selection is still proceeding in humans due to our technology, there is a short answer: Speed is probably not very important in natural selection and certainly it is not the only consideration.
There is a danger in mutation. Most mutations do not help the species survive. A species and an environment exist in balance with each other. Populations simply adapt to their current surroundings and the changes to them; they do not necessarily become better in any absolute sense over time. A change in the environment may require a change in the species in order for it to survive. But if a mutation spreads too quickly across an entire species it may prove maladaptive to the species if the environment undergoes a further change.
More diversity in mutations and hence change is probably better than speed in a mutation becoming widespread in a species.
There is also the possibility that a trait that is successful at one time may be unsuccessful at another. This principle was demonstrated by the classic experiments by Drs. C. Paquin and J. Adams. They developed a yeast culture and maintained it for many generations. Every so often, a mutation would appear that allowed its bearer to reproduce better than its contemporaries. These mutant strains would crowd out the formerly dominant strains. Samples of the most successful strains from the culture were taken at various times.
In later competition experiments, each strain would outcompete the immediate previous dominant type in a culture. But interestingly, some earlier strains could outcompete strains that arose later in the experiment. Competitive ability of a strain was always better than its previous type. Yet competitiveness in a general sense was not increasing. The success of any organism depends on the traits of its contemporaries. There is likely no optimal design or strategy for most traits, only ones based on chance such as the competition and the environment.