More than Moore: Comparing forecasts of technological progress
8:00-8:20 am, October 4
An intuitive way to illustrate technological change is to plot some performance metric as a function of time, such as Moore's law (and its various generalizations). However, we should be careful with the interpretation of such plots because the shape of the trend crucially depends on which metric we choose.
Furthermore, time may not be the best predictor of future performance, calling into question the validity of such trend projections. We compared the historical prediction accuracy of different functional forms for performance curves, based on a number of information technologies and energy technologies. Our results suggest that taking into account production data may lead to better forecasts than relying on time alone.
This research highlights the need for a comprehensive Performance Curve Database to document technological trajectories and to enable identifying general patterns of change and modeling the dynamics of innovation. A preliminary prototype is already under development at the Santa Fe Institute with online access at http://pcdb.santafe.edu accepting uploads for visualization, comparison, and prediction.
Biographies: Béla Nagy
Béla Nagy is a quantitative futurist, advocating the development of a data-driven, predictive science for understanding and modeling scientific and technological progress. Other interests include extreme life-extension (Calorie Restriction), long-term equity investments, and teaching future-oriented thinking (personal coaching).
After his undergraduate studies in Math and Computer Science at Eötvös Loránd University and the University of Waterloo, Béla worked on a wide range of software R&D projects. Experimenting with cognitive algorithms for an AGI developed by Adaptive A.I. led him to do a PhD in the Design and Analysis of Computer Experiments at the University of British Columbia, where he designed and implemented Fast Bayesian Inference to quantify the uncertainty in the predictions of computer models.
He is currently researching Technology Evolution at the Santa Fe Institute and is developing a Performance Curve Database at http://pcdb.santafe.edu to facilitate collaborations between research institutions, think tanks, and other parties in the private and public sectors interested in technological innovation. The goal is to collect a large body of empirical trends that will enable this community to assess and compare the historical track records of various forecasting procedures and devise new methods for technology foresight and management.
