Donald Campbell — On AI
This page lists every Orange Pill Wiki entry hyperlinked from Donald Campbell — On AI. 17 entries total. Each is a deeper-dive on a person, concept, work, event, or technology that the book treats as a stepping stone for thinking through the AI revolution. Click any card to open the entry; in each entry, words colored in orange link to other Orange Pill Wiki entries, while orange-underlined words with the Wikipedia mark link to Wikipedia.
Blind Variation and Selective Retention
Campbell's 1960 thesis that all knowledge acquisition — from amoeba to scientist — operates through the same mechanism: generate possibilities not directed by foreknowledge of the solution, then selectively preserve those that prove valuabl…
Campbell's Law
The principle — stated by Campbell in 1976 — that any quantitative indicator used for decision-making will be corrupted by the pressure it creates, and will distort the process it was intended to monitor.
Flow State
Mihaly Csikszentmihalyi's name for the condition of optimal human engagement — and, in Wiener's framework, the subjective signature of a well-regulated negative feedback system.
Next-Token Prediction as Universal Objective
The single training objective — given a sequence, predict what comes next — that, scaled sufficiently, produced almost everything a frontier language model can do. The most consequential methodological discovery of the AI decade.
Tacit Knowledge
Michael Polanyi's term for the knowledge that lives in the hands and nervous system rather than in explicit propositions — acquired through practice, failure, and embodied pattern recognition, and dissolved by AI workflows that produce ou…
The Beaver's Dam
The canonical example of allogenic ecosystem engineering — a structure that modulates rather than blocks the flow of its environment, creating the habitat pool in which diverse community life becomes possible.
The Convex Hull of Knowledge
The geometric frame, drawn from computational geometry, for understanding what AI-generated outputs can and cannot reach — the space defined by weighted combinations of the training corpus's statistical regularities.
The Interpolation Trap
The structural phenomenon by which AI-generated outputs trigger the human experience of discovery while actually performing sophisticated retrieval — the most dangerous failure mode of smooth, plausible, rhetorically effective machine outpu…
The Productivity Number
Edo Segal's twenty-fold multiplier from Trivandrum — received by the culture with the reverence a quantitative civilization reserves for quantitative claims, and the archetypal thin description of a transformation whose meaning lives elsew…
The Selective Retention Function
The domain-specific capacity — built through years of direct engagement with a field's resistance — to distinguish the significant anomaly from meaningless noise, and the scarce cognitive resource on which the entire value of human-AI colla…
Vicarious Selectors
Campbell's term for mechanisms — vision, language, culture, AI — that perform the trial-and-error of lower-level processes vicariously, at lower cost and higher speed, by trading breadth of search for efficiency of search.
Edo Segal
Serial entrepreneur and technologist whose The Orange Pill (2026) provides the phenomenological account — the confession over the Atlantic — that Pang's framework diagnoses and treats.
Norbert Wiener
American mathematician (1894–1964), founder of cybernetics, and the first public voice warning that automation would pose new kinds of governance problems.
Fleming's Penicillin
The 1928 laboratory accident in which an open window, an airborne mold spore, and a bacteriologist's trained retention function produced the most important class of drugs in medical history.
Röntgen's X-Rays
The 1895 discovery — in which a fluorescent screen glowed faintly across a Würzburg laboratory — that demonstrated selective retention function specificity with experimental clarity.