Pamela McCorduck

In the [YOU] on AI Field Guide

The cycle that began with [YOU] on AI insists on locating the present AI transformation within the broad sweep of human history rather than the frenzy of the news cycle. McCorduck is the historian who most completely accomplished this location, and her work is the cycle's deepest source for the intuition that what is happening now is continuous as well as unprecedented. Her central thesis—that the wish to forge the gods is not marginal but central to human culture, recurring across civilizations and centuries with the same hopes and the same anxieties—provides the temporal frame within which the cycle's more immediate concerns make sense. The vertigo that users feel before a chatbot that seems to understand them is not a new feeling; it is the oldest feeling human beings have about their own ingenuity, and McCorduck gave us the evidence for that continuity.

She also modeled the intellectual posture the cycle most needs and most rarely achieves: the refusal of both poles. She was not a cheerleader, though she loved the field and the people in it with real warmth. She was not a doomsayer, though she lived long enough to regret, openly, that she had not warned sooner. She held both truths at once, understanding that the technology was genuinely powerful for good and that its goodness was not guaranteed, that it had to be fought for through deliberate choices. The cycle's own stance—the amplifier carries whatever signal is fed into it, and the quality of the signal is the human's responsibility—is a restatement of the technological optimism she confessed and then complicated.

Her late regret has specific force for the cycle's readers. McCorduck named deepfakes specifically as the abuse she wished she had foreseen and flagged: the decorrelation of fluent presentation from truth, the manufacturing of convincing falsehood at scale, the erosion of our ability to trust what we see and hear. These are exactly the concerns the cycle develops through the lens of the present AI moment. She saw the outline of the problem in 2019; she left the warning as a gift to those who came after. Taking that gift seriously is part of what it means to read her.

Origin

Born in Liverpool in 1940 during the Blitz and raised in California, McCorduck came to AI almost by accident. As an undergraduate at Berkeley, she helped type and organize the materials for what became Computers and Thought, the foundational 1963 anthology assembled by Edward Feigenbaum and Julian Feldman. She was, almost by accident, present at the creation. Unlike nearly everyone else in the room, she was equipped to notice what kind of creation it was: not the birth of a new engineering discipline but the materialization of a three-thousand-year-old human wish.

Through her marriage to the computer scientist Joseph Traub and through decades of patient relationship-building, she earned entry to the small world that invented the field. She interviewed Herbert Simon and Allen Newell at Carnegie Mellon, sat with John McCarthy and Marvin Minsky, knew Herbert Simon and Marvin Minsky well enough to render them not as icons but as characters—with their hubris and their doubts and their particular ways of being wrong. Machines Who Think, published in 1979 and expanded across twenty-five years, is the result: the indispensable account of where AI came from, indispensable not because it is comprehensive but because it is alive.

She went on to co-write The Fifth Generation with Edward Feigenbaum in 1983, sounding an alarm about Japanese AI ambitions that was enormously influential but ultimately misjudged the paradigm's staying power—a misjudgment she was honest enough to revisit and learn from. She later wrote The Universal Machine, Aaron's Code (a study of machine creativity), and The Futures of Women (with Nancy Ramsey). She died in 2021, on the eve of the large language model moment she had spent a lifetime preparing to understand.

Key Ideas

The ancient wish to forge the gods. McCorduck's foundational claim is that the wish to make a thinking artifact—to externalize and replicate the mind—is as old as human storytelling and as culturally central as any religious impulse. She traced it through Talos, the Golem, Pygmalion, Hephaestus, and the clockwork automata of the Enlightenment, arguing that these were not idle curiosities but rehearsals: humanity imagining the thinking machine long before it could build one, the way a body prepares for a motion it has not yet made. The same hopes recur (the mind that serves and transcends us), the same fears recur (the creation that escapes our control), and the engineers of the twentieth century inherited both, fully formed, encoded in stories their culture had been telling for millennia.

AI as mirror. Making a mind requires deciding what mind is. Every technical paradigm in AI has therefore carried within it an anthropology—a picture of the creature being imitated. When the founders modeled the mind as symbol manipulation, they were proposing a vision of the human as logician. When later researchers turned to neural networks and learning, they proposed a different human, more pattern-matcher than logician. McCorduck saw this double movement—the building of AI as simultaneously a technology and a theory of human nature—with a clarity no insider could achieve. The failures of AI she found as instructive as its successes, because every time a machine could not do something humans found easy, it revealed a hidden competence the builders had taken for granted.

Meta-art and machine creativity. In her 1991 study of the painter Harold Cohen and his generative program AARON, McCorduck confronted the question that the rest of AI preferred to avoid: can a machine be creative, and if it can, what becomes of creativity's claim on human specialness? She refused both cheap answers: AARON was not merely Cohen with extra steps, because it produced images Cohen did not anticipate; but it was not an autonomous artist, because it had no experience of beauty and no stake in the work. Her resolution was the concept of meta-art: the art was no longer only the drawing on the wall but also the system that made the drawing, and building such a system was itself the deeper creative achievement. The locus of creativity had shifted up a level—from artifact to generator of artifacts.

The cyclical structure of AI hype. McCorduck lived through enough AI winters and springs to map the cycle from the inside. The Japanese Fifth Generation program she helped sound the alarm about in 1983 collapsed; the expert systems boom she contributed to went bust; the neural network revival came and went and came again. She came to see the pattern not as failure but as the characteristic rhythm of a genuine and difficult project. The wish does not fail; the specific paradigm fails, goes underground, and returns in unexpected form. Her temporal wisdom—the ability to distinguish the permanent transformation from the current froth—is among her most valuable bequests to the present moment.

The courage of late regret. In her final memoir, McCorduck expressed regret that she had not warned sooner about the technology's potential for misuse. This confession took real courage: it would have been easy to let the memoir be a victory lap. Instead she used it to audit her own optimism, acknowledging that her hopeful framing might have left readers underprepared for the darker uses the technology would find. The technological optimist had seen enough to know that optimism alone was insufficient. She named her failure and left the naming as a gift.