machine learning – What does the decision boundary of XOR problem look like?

My textbook walks through an example of solving the XOR problem in machine learning using a two-dimensional RBF network. It does this by setting the centers for the two basis functions at (0,0) and (1,1). Afterwards, it challenges us and asks to imagine what the decision boundary for an RBF would look like if the two centers were instead at a sample of each class (class 0 could be (0,0) and class 1 could be (1 0)). Since there is only two classes, wouldn’t this result in essentially a straight line, or would it look like something else? Is there any reason why we would want to do this?