Classful addressing divides IP addresses into classes A, B, C, D, and E, limiting flexibility and wasting addresses. Classless addressing, also known as CIDR, provides more flexibility by allowing companies to request the specific number of addresses needed rather than being restricted to block sizes. CIDR network addresses include a subnet mask to specify the network and host portions of the address, allowing discontiguous networks to be created from a classful address range.