This paper expands on Robert M. Solow's 1957 paper and develops an aggregate production function with perfect goodness-of-fit that can be applied to any types of data, but demonstrates ultimately that there is no such thing as an aggregate production function. Solow's theory can even be expanded to scenarios without the assumption of constant-returns-to-scale. However, input coefficient estimation turns out to be self-fulfilling. The production function can take any form, Cobb-Douglas or others, linear or non-linear. Technology is only a means to eliminate estimation errors