I.J. Good's 1965 'ultraintelligent machine' paper originated the intelligence-explosion idea

Irving John Good — a Bletchley Park cryptanalyst who worked alongside Alan Turing in Hut 8, and later co-coined the "Bayes factor" — is the same statistician who authored the 1952 logarithmic scoring rule now used as the LLM training loss. In 1965, in "Speculations Concerning the First Ultraintelligent Machine," he defined the ultraintelligent machine as "a machine that can far surpass all the intellectual activities of any man however clever" and argued that, since designing machines is one such intellectual activity, such a machine could design still better machines in a runaway feedback loop. He called it the "last invention that man need ever make" — the founding statement of what is now termed the intelligence explosion.

The claim places a single person under two pillars of modern AI: the calibration objective that trains the network, and the origin of the superintelligence argument that motivates AI-risk discourse. Good's framing is the ancestor of the contemporary existential-risk register — the same rhetorical territory claim-yao-frames-ai-as-extinction-level-new-species occupies, and the optimistic pole against which the 1970s AI contraction reacted (claim-lighthill-1973-named-general-problem-solving-disappointing). Good's cultural footprint is also direct: he consulted on Stanley Kubrick's 2001: A Space Odyssey, the film whose HAL 9000 is documented separately in claim-clarke-hal-9000-death-song-from-bell-labs-demo.

His 1965 optimism did not hold; he reversed it decades later (claim-ij-good-1998-reversed-intelligence-explosion-to-extinction).