I.J. Good

I.J. Good (1916–2009) was a British mathematician and statistician — a Bletchley Park cryptanalyst under Turing, later a statistician at Virginia Tech — whose 1965 essay coined "intelligence explosion" and the "ultraintelligent machine," and whose 1998 reflection reversed his own earlier optimism into a warning that such a machine could bring human extinction. His 1952 logarithmic scoring rule, developed decades earlier in an unrelated context (calibrated probability elicitation), turns out to be mathematically identical to the cross-entropy loss now used to train large language models — a genuine cross-domain bridge, not mere trivia. He matters to the vault as the earliest anchor for the intelligence-explosion-to-extinction lineage that later AI-safety rhetoric, including Andrew Yao's 2024 "new species" warning, echoes decades later from an unrelated field.

References

• claim-ij-good-1965-ultraintelligent-machine-coined-intelligence-explosion — the 1965 coinage
• claim-ij-good-1998-reversed-intelligence-explosion-to-extinction — the 1998 reversal to existential risk
• claim-ij-good-1952-logarithmic-score-is-the-llm-cross-entropy-loss — the 1952 scoring-rule/cross-entropy bridge
• Related: claim-yao-frames-ai-as-extinction-level-new-species — a later, unrelated-field echo of Good's extinction framing
• Captures: 2026-07-12 (original three notes); 2026-07-27-does-the-carnegie-endowment-piece-confirm-cnaisdas-bilingual (flagged Good as an entity candidate — hub built this session)