The style is never entirely separate. Putting a class "side-image", "inline-image" or "centered-image" is a styling choice. Sure you put a lot of the detailing stuff in the css, but you can as well trivially define your own macro in Latex to separate most of the gnarly stuff, or define your own style.
If you don't want to be able to control precisely individual elements, then there's only semantic markup left. I would say the main difference between the two is that in Latex it requires slightly more knowledge to use good practices. In html there's no difficulty difference between good and bad practices, but there's nothing to enforce good practices either.
That's why proponents of "semantic css" (RIP) would have said the correct classes would have been "secondary-image" or "main-image" or "accent-image" or something like that, with the stylesheet deciding if if "secondary-image" should be at the side, alternating, or whatever.
