What has agents research accomplished and where do we need to go? I will answer these questions by telling the story of two softbots:  Rodney (1991 - 1997) is an early softbot that relied on a declarative action representation and on a novel algorithm for planning (and acting) with incomplete information to satisfy a wide range of high-level user requests;  Jango (http://jango.excite.com) is a popular shopping assistant able to access hundreds of on-line stores in over fifteen product categories. Both softbots enable a person to state "what" she wants accomplished, the softbot is responsible for "how" and "where". Their limitations suggest directions for future work for both academia and industry.

In this talk I will rejoin a discussion in which I first participated ten years ago, on the question of machine agency.  My renewed interest in questions of agency is inspired by developments over the last decade both in the area of interactive computing and in the debate over agency within social studies of science and technology.  What I propose to do is another attempt at working these two projects together, in what I hope will be a new and useful way.  Drawing on recent discussions within technology studies, I will argue that agency is locateable neither in humans nor in machines, but only in relations between them.

Lucy Suchman received a Ph.D. in Social/Cultural Anthropology from the University of California at Berkeley in 1984.  Since 1979 she has been a researcher at Xerox Palo Alto Research Center, where she is currently a Principal Scientist and heads the Work Practice and Technology area.  Her research concerns the relation of everyday working practices to computer systems design.  Her dissertation "Plans and Situated Actions: the problem of human-machine communication" was published by Cambridge University Press in 1987.  It provided a critical analysis of models of human action and communication that underlie the design of interactive machines, and proposed an alternative perspective drawn from recent developments in the social sciences.

We use computers to study economics, but few people realize that we can use economics to study and design computational systems. The reason is that  agents roaming in computer networks can be regarded as a community of processes that in their interactions, strategies and lack of perfect knowledge, face the same issues as people in markets. This talk will  describe a number of economics approaches that are useful for designing agent systems and for understanding their dynamics, with examples ranging from from electronic transactions in cyberspace to surfing the Web.

Bernardo A. Huberman is a Research Fellow at the Xerox Palo Alto Research Center, where he heads a group involved in studying the dynamics of Internet ecologies. He received his Ph. D. in Physics from the University of Pennsylvania, and is concurrently a Consulting Professor of Physics and  of the Symbolic Systems Program at Stanford University. He has worked in condensed matter physics and the theory of critical phenomena, and is one of the discoverers of chaos in a number of physical systems.  He also established the existence of a number of universal properties in nonlinear dynamics, and his research into the dynamics of complex systems led to the discovery of ultradiffusion in hierarchical structures.

For the past nine years, he has been working in the area of economics and computing with enphasis on  resource allocation in distributed computational systems. He has recently developed a portfolio method for solving hard computational problems, and for speeding electronic  transactions in the Internet.  He started the field of ecology of computation, and is the editor of a book on the subject.  His team designed and implemented Spawn, a market system for the allocation of idle resources among machines in computer networks, and a thermal market mechanism for the control of building environments.

Recently, Dr. Huberman has been studying the power of cooperation in collective problem solving by many agents, and the dynamics of Internet storms. He coauthored an article on  Social Dilemmas in the March 1994 issue of Scientific American.Dr. Huberman is a Fellow of the American Physical Society, a former trustee of the Aspen Center for Physics and Fellow of the Japan Society for the Promotion of Science. He is the co-winner of the 1990 CECOIA prize in Economics and Artificial Intelligence, and this year he received the IBM Prize in Computing and Finance at the third International Conference on Computational Economics.  He has held visiting professorships at the University of Paris, France, and the University of Copenhagen, Denmark.