My understanding is that the mutation rate has evolved (see
Darwin in the Genome by Lynn Caporale).
I beleive that Caporals is correct. As the mutation rate decreases, so does the fitness advantage of further "improvements" in copying fidelity. At some point we hit the situation where the mutatiion spreads so slowly that environmental changes will arrive before it becomes fixed in the population.
It isn't a plausible explanation for the dinosaur extinction because that would require many different species to suffer the same effect, independantly.r