Both art and entertainment are taking a strong interest in computer-generated agents that act autonomously. Several challenges face their creators. Since a large portion of the computational resources in interactive works is devoted to graphics, where and how do we locate the computing power required to drive artificial intelligence? As autonomous action competes with displayed behavior for processor clock cycles, what can one do to graphics to make them both more efficient and more expressive? One solution is to prerender imagery and find better ways to manage it. Another challenge one faces in acquiring the ability to give birth to interesting creatures is finding a suitable tool box. We need to develop intuitive interfaces that make it possible for designers to craft individual personalities without the need to resort to programming. This is a key in making the creative process painless and efficient enough to generate a variety of healthy progeny.