This should have said, that nearly all they told has not really anything to do with Git branches, but with a methodology to use them. The only Git concept that comes close to the concept of inheritance relations of branches is the first-parent, second-parent, <n>th-parent thing, something they haven't even mentioned.

A branch in Git is really just an auto-increasing tag, with an attached remote. Telling people it is some part of a history or that there is any ordering relation between different commit ranges with the same ancestor is only setting them up for failure. This seems like yet another article from the "Git is Github" crowd.