Turing's 1950 paper was far stranger than the 'Turing Test' it's remembered for — he seriously entertained ESP, proposed building child-minds instead of programming adult ones, and set a prediction for the year 2000 that roughly came true in the most disputed way possible. His death in 1954 left a conceptual vacuum, and the field that emerged in his absence lurched between grand promises and supposed 'winters' that were partly mythological — US AI society membership actually tripled during the first one.