Radnet, Inc.

The Long Range Future of Computation: Possibilities and Limitations

During the last 50 years, computation per dollar has doubled every 2 years. This progress has continued through various technology changes (from mechanical computers to relays - vacuum tubes - transistors - simple integrated circuits - LSI - VLSI ....). Without knowing exactly how future computers will be constructed, we can still examine various physical constraints to see how much further the current rate of progress can continue. This subject has been of great interest since such progress was first noted some 40 years ago.

The nearly universal consensus of various papers and talks given on the subject during the past 40 years has always been the same. No matter when the question has been addressed the conclusion seems to have always been that the progress in computation per dollar must come to a stop in about 10 more years (from the date of the paper!). Our conclusion is very different. We will explain why it is quite likely that the same rate of progress will continue for more than an additional 100 years and also why and when that rate of progress must absolutely come to a stop. As an added bonus we will describe a simple algorithm that allows for the solution, in polynomial time, of all very large (but not too large) exponentially difficult computations.

Ed Fredkin has worked with a number of companies in the computer field and has held academic positions at a number of universities. He is a computer programmer, a pilot, advisor to businesses and an amatuer physicist. His main interests concern digital computer like models of basic processes in physics.

Return to Inventing the Future Home Page