The grid usage is pretty weird. It’s essentially using grids to implement an 8px base unit, which can be achieved with less complexity by using css variables. Would make a bit more sense if it used flexible units.
Given the hard coded amounts used in the CSS presented in this article (eg 11, 13), they must be using some layout framework that translates their language into proper grid positions so that must be why the grid is used in such a way.
I had the same thought but I don’t think css variables are necessarily better for this use case. It’s simple to position and reposition elements in a grid depending on media queries, and is it any less efficient when rendering than css variables?