Frictionless forecasting is a fiction
With the increase in specialization, experts are available today who are capable of short-term forecasts about almost all aspects of social and economic behavior.
Concurrently, it is becoming increasingly difficult for any one person to integrate all of this understanding.
Computer technology now allows one to incorporate this immense body of short-term forecast knowledge in programs that can, through simulation, predict the long-term future impacts of contemplated decisions. These impacts are often surprising and counterintuitive, accounting for the fear, in the past, of anything but small changes2.
Once we can predict the impact of globally discontinuous decisions, debate can appropriately center on which effects are preferable and not on what the impacts will be.
KeywordsInference Rule Incremental Change Synaptic Modification Advance Information Technology Situational Aspect
Unable to display preview. Download preview PDF.
- 3.Werbos, Paul J., “A menu of designs for reinforcement learning over time,” in Neural Networks for Control, T. Miller, R. Sutton, and P. Werbos, eds., MIT Press, Cambridge, MA., 1990, 67–95Google Scholar
- 4.Simon, Herbert A., “The Mind’s Eye in Chess,” (with W.G. Chase), in Models of Thought, Yale University Press, New Haven, CT., 1979, 421. See also his Hitchcock Lecture delivered at the University of California at Berkeley, Feb. 13, 1990.Google Scholar
- 5.A more detailed exposition of our skill-acqusition model, directed more toward coping skills than predictive skills, may be found in our book: Dreyfus, Hubert L. and Dreyfus, Stuart E., Mind over Machine, paperback edition, Free Press, New York, 1988Google Scholar
- 6.Dreyfus and Dreyfus, Mind Over Machine, op cit, 36–40, 163–167Google Scholar