I first became interested in the idea of centralized economic planning while I was an undergraduate computer science major at Worcester Polytechnic Institute in the early seventies. I first went to college thinking I’d be an astronaut or something. I loved astronomy and math (you know, the moon landing and all) so I thought I would major in physics. I had this great professor in a my first computer science class and really got into computer programming. In particular I took an interest in the emerging field known as “artificial intelligence.” The combination of exposure to the natural sciences and to AI drew me to an interest in philosophy, especially the philosophy of knowledge. I studied philosophers such as Hubert L. Dreyfus, (see his  1992 What Computers Still Can’t Do: A Critique of Artificial Reason, revised. Cambridge, Mass.: MIT Press) who clarified for me what computers are not good for. In terms of creative thinking or the use of language computers are still pretty stupid and they were outrageously dumb in the 1970’s.
That was when I ran across reference to a famous Polish economist named Oskar Lange, who had remarked that had he only known of the possibilities of electronic computers he would have been able to more easily answer Ludwig von Mises’s famous critique of centralized planning. Mises had argued that the separation of ownership into the hands of different and possibly contentious agents serves a knowledge-generating function in the economy, putting information about the relative scarcities of different goods into the price system. Lange in the 1930’s responded by concocting a model of “market socialism” whereby the price system is replaced by a central planning board that replicates the welfare ideal according to the principles of neoclassical economics: marginal cost pricing. Now that we know of computers, he said, Mises’s challenge is more easily disposed of. The computers could directly calculate the solutions to the general equilibrium economists’s equations, and so we could replace the market!
This looked like weird stuff to me. The market process seems like the kind of thing that inherently involves the complex interaction of a variety of different persons, each tugging and pulling in different directions. I presume that the participants to market activities bring their full intellect to bear on their thinking. Replicating the outcome of the interaction of a multiplicity of minds sounded like AI2 . If AI seems mostly to be overambititous, this Lange guy had to be way off.
So I started looking into this original debate, and found the work of Ludwig Mises. The view of knowledge in economic processes that I picked up from Mises and some of his followers, such as Friedrich Hayek, Ludwig Lachmann, and Israel Kirzner, has been on my mind ever since. My first two books were aimed at articulating Mises’s critique of centralized planning, the “calculation argument,” and in a number of subsequent articles I have kept coming back to different aspects of this idea.