In 1943 Warren McCulloch and Walter Pitts published the article that founded artificial neural networks. They demonstrated that formal neurons — simple binary units that activate when the sum of their inputs exceeds a threshold — could be organized into networks capable of computing any logical function. The implication was radical: computation and nervous activity share a common mathematical structure. Pitts, only 19, had been essentially self-taught and was homeless when he wrote the article. The McCulloch-Pitts model does not learn — its weights are fixed — but it established the conceptual architecture on which all later neural networks would be built.