The Perceptron

Frank Rosenblatt at Cornell builds the Perceptron, a learning machine that adjusts its own weights in response to training examples. The Navy funds it; the press announces it as a machine that will soon walk, talk, see, and be conscious of its own existence.

Marvin Minsky and Seymour Papert’s 1969 book Perceptrons demonstrates that single-layer networks cannot learn certain logical functions, triggering a collapse in enthusiasm. Funding dries up. The first AI winter begins.

The Perceptron episode establishes a pattern that recurs throughout AI history: a genuine capability, inflated claims, institutional investment, discovery of limitations, withdrawal of support.

Sources: Frank Rosenblatt, “The Perceptron: A Probabilistic Model for Information Storage and Organization in the Brain,” Psychological Review, 65(6), 1958, 386–408. Marvin Minsky and Seymour Papert, Perceptrons: An Introduction to Computational Geometry (MIT Press, 1969).