n July 1958 the United States Navy showed reporters a machine that learned. The newspaper account is a primary document twice over: of the machine, and of how people talked about learning machines from the first day.

the document

(The New York Times, 1958). A short wire report on page 25. It opens:

“The Navy revealed the embryo of an electronic computer today that it expects will be able to walk, talk, see, write, reproduce itself and be conscious of its existence.”
— The New York Times, 8 July 1958

The demonstration itself was modest. The program ran on an IBM 704, a room-sized general-purpose computer, and the report says it learned to tell left from right after fifty trials. The gap between those two paragraphs is the subject of this page.

the paper

The scientist was Frank Rosenblatt, a psychologist at the Cornell Aeronautical Laboratory in Buffalo. His paper that year (Rosenblatt, 1958) framed the perceptron as a theory of how a brain might store and organize information, with connections strengthened or weakened by experience. A purpose-built machine, the Mark I Perceptron, followed, and the full theory filled a book (Rosenblatt, 1962).

what happened next

Minsky & Papert proved what a single layer of adjustable connections cannot compute, such as whether a figure is connected. The book is often blamed for a long cooling of interest in neural networks; the history is more tangled than that, but the limits were real, and multi-layer networks were the way around them.

Why it lasts. The pattern of a modest demonstration and an enormous forecast has repeated with nearly every machine that answers. It is a lesson in anthropomorphism and in mental models: what people believe about a system is shaped by how it is described to them.

in the tour: the room of machines that learn has a placard called "The perceptron, 1958" ···

References

1. The New York Times (1958). New Navy device learns by doing; psychologist shows embryo of computer designed to read and grow wiser. 8 July 1958, p. 25 (a UPI wire report from Washington, 7 July). www.nytimes.com/1958/07/08/archives/new-navy-device-learns-by-doing-psychologist-shows-embryo-of.html
2. Rosenblatt (1958). The perceptron: A probabilistic model for information storage and organization in the brain. Psychological Review. doi:10.1037/h0042519
3. Rosenblatt (1962). Principles of Neurodynamics. Washington, DC: Spartan Books.
4. Minsky & Papert (1969). Perceptrons. Cambridge, MA: MIT Press.

see also