The stored-program computer

On 30 June 1945 the Moore School circulated a 101-page typescript with John von Neumann's name on it. It described a machine that kept its instructions in the same memory as its data, so that a program could be loaded, changed and even rewritten by the machine itself. Every general-purpose computer since has followed that plan, and the phrase "von Neumann architecture" stuck, to the irritation of the engineers Eckert and Mauchly who had built much of the thinking with him.

The report matters here for a smaller reason. When von Neumann needed a notation for the machine's logical elements, he did not draw vacuum tubes. He drew McCulloch–Pitts neurons, citing the 1943 paper directly and treating the computer's gates as idealised nerve cells. The first design of the modern computer was written in the language of a brain model.

That habit of mind shaped the field. The people who built the first machines already believed that computation and thought were the same kind of thing, and that a sufficiently large network of simple switches could do what a nervous system does. Within a decade Dartmouth would make the belief a discipline.