IMO, Python 3's support for matrix multiplication using the @ operator is itself worth the cost of admission. Much of technical computing is just implementing algorithms that use linear algebra extensively, and if you're coming from matlab littering your code with dot(dot(X,Y),Z) is a real pain