I like the analogy that biologists are making code changes (especially with genetic therapies) without actually understanding the machine code specification or even having a copy of the source code.

It's like a hacker flipping bits in a binary trying to figure out what's going to happen.. except the hacker at least can look up the complete machine code.