The expert system/artificial intelligence portion of the molecular graphics system has been the slowest to get started. Francois Major brings to the project a knowledge of LISP and computer reasoning. We are in the process of writing rules about how proteins fold using the insights from UPATSU. The goal is to provide high-level recognition and guidance routines that will enable us to build models of proteins faster than can be achieved by calculational methods alone. We anticipate a convergence when in the next generation workstation we will be able to have the four requisites (graphic, computation, control and intellection) in the same machine.