Changing a single bit in computer code is very likely to have an arbitrarily large, commonly fatal effect. Changing a single base in a gene rarely does. But a lot of bases changed is likely to have some effect; the more changes, the more effect.

What critics of natural selection always miss is that the set of small changes between one state and another, in a breeding population, do not need to happen one after the other. It could take hundreds of steps, but they all may exist scattered through the population at the same time, and can over just a few dozen generations propagate to all, when conditions favor that; or even fewer if lacking most is fatal.

Bacteria need a lot more generations to evolve, because they don't swap genes around much. But their generations are very short. Some mutations they do swap around, particularly antibiotic resistance, and ability to infect a new host, via plasmids. Anthrax normally lives in soil, but picks up virulence genes from plasmids it happens upon.

The creative power of randomly mutated code is, again, easily verified.