So a protein is a very big molecule. It’s so big that parts of it are attracted to other parts and cling together, think of a balloon you’ve rubbed on your hair and stuck to a wall.

These self-stickiness gives proteins a folded shape. There’s still some charge, but since the protein is folded a specific way, most other molecules don’t fit to bind to those charged areas.

However, most proteins have at least one other protein that does fit. So when a protein “meets its match” the two can bind together. When they do that, their charges mingle and change and the proteins change shape.

This act, proteins changing shape as a result of things binding to them, is how your cell “does” stuff.

So in your Golgi apparatus example, the proteins don’t know what they’re doing at all. But they have a specific shape and charge and so attract and bind specific other proteins. When they do so, they change shape and that results in the golgi doing something.

Eh. It’s hard to overstate how much of the genome isn’t code-specific. That is, some of it is useless, some of it is structural (need x bases of any sort), some of it is compositional (need a lot of g and c but the precise ratio isn’t important) etc

A lot of the major protein-coding, structural, and regulatory stuff is highly conserved, so there’s a lot of overlap between any two species (Eg humans + bananas)

But all of that other stuff? Eh. It can vary basically as much as it wants consequence-free, producing a lot of within species differences.