In 1969 Marvin Minsky and Seymour Papert, at MIT, publish "Perceptrons", mathematically proving the single-layer perceptron cannot solve non-linearly-separable problems (e.g. XOR). The rigorous demonstration devastates neural-network research funding for over a decade — the first AI winter. Minsky and Papert knew multi-layer networks could overcome this but claimed training them would be computationally intractable — correct about single-layer limits but wrong to not anticipate backpropagation (Rumelhart, Hinton, Williams, 1986).