The technological singularity hypothesis asserts that the invention of a synthetic intelligence with greater cognitive capacities than a human being will trigger an exponential increase in synthetic cognition and knowledge. Each generation of synthetic intelligences will be able to create new generations of cognitive beings with even greater capabilities than themselves. Some projections envision a future with superintelligences with millions or billions times the cognitive capability of human beings. This paper will argue that the primary function of cognition is to predict the future and make plans based on those predictions. Exponential increases in cognitive capability and knowledge do not necessarily result in exponential increases in the ability to predict and plan for the future. We will show that exponential increases in knowledge may only result in modest linear increases in the ability to predict the future or make plans. Therefore, the actual capabilities of superintelligent machines may only be minimally, if at all, greater than current human capabilities.
展开▼