Lex Fridman PodcastPamela McCorduck: Machines Who Think and the Early Days of AI | Lex Fridman Podcast #34
CHAPTERS
Why McCorduck wrote 'Machines Who Think' (and why it was harder than expected)
Lex introduces Pamela McCorduck and the mythic framing of AI as an ancient human urge. McCorduck explains how she moved from novelist to historian-by-necessity, deciding to interview AI researchers directly and discovering the work was far more demanding than her initial plan.
Meeting the founders while they were still building the field
McCorduck describes what it was like to interview the living founders of AI and how seriously they took their own importance. She reflects on the generosity of these researchers at the height of their careers, and how unusual it was to capture a field in real time.
Dartmouth and the “four founding fathers,” plus the Logic Theorist milestone
The conversation zooms in on the Dartmouth era and the personalities who shaped it. McCorduck identifies the four “founding fathers” and highlights how Newell and Simon arrived with an actual working program—Logic Theorist—while others had mostly ideas.
What drew McCorduck to AI: intelligence outside the human skull
McCorduck explains the central fascination that pulled her into the field: the possibility of non-human intelligence. She contrasts her mythic framing with the founders’ pragmatic motives and notes how AI reshaped what scientists consider “intelligence.”
Ancient roots: Homeric robots, early automatons, and Babbage & Lovelace
McCorduck traces AI’s lineage through myths and early technological aspirations long before computers. She points to robots in Homer, early 20th-century attempts at automata, and Babbage/Lovelace’s recognition that computation could imply intelligent behavior.
Hellenic vs Hebraic attitudes toward robots—and the blasphemy fear
McCorduck offers a cultural lens: Greek (Hellenic) traditions welcoming helpful robots versus Hebraic traditions wary of human imitation and idolatry. This becomes a way to explain why AI often triggers moral panic framed as blasphemy rather than practical risk.
Funding struggles and being dismissed as “only a writer”
McCorduck recounts how difficult it was to secure institutional funding for a history of AI. After rejections from major agencies, she ultimately receives private support from Ed Fredkin—who liked “crackpot ideas”—enabling the project to happen.
CMU life, AI as outcast discipline, and Don Knuth defending AI’s place
The conversation shifts to the academic positioning of AI relative to computer science. McCorduck describes how AI was treated as marginal at times, including an episode where computer science leaders wanted to exclude AI—until Don Knuth insisted it mattered.
Friendship with Herb Simon: sherry, salons, and serious conversation
McCorduck shares personal stories about becoming close friends with Herbert Simon. Their relationship centered more on literature, music, art, and ideas than technical AI, and it led to creating a salon-style discussion group with Newell, Simon, and others.
Symbolic AI vs deep learning apps—and why “AI winter” is a misleading story
McCorduck reflects on how AI evolved in ways she didn’t predict: modern success is heavily algorithmic and data-driven, while symbolic reasoning feels underdeveloped. She argues “AI winter” is largely a commerce narrative—basic research continued, but monetization lagged.
Hype, funding, and the UK’s Lighthill Report: how policy can stall science
Lex and McCorduck discuss how over-promising can trigger backlash, using autonomous vehicles as a modern example. McCorduck emphasizes that while science can be harmed by funding cuts, the deeper pattern is tension between commercial timelines and scientific reality, illustrated by the UK’s Lighthill Report and later DeepMind resurgence.
Singularity skepticism and a more gradual view of machine intelligence
McCorduck pushes back on Kurzweil-style “game over” singularity narratives, arguing machines have long surpassed humans in narrow tasks. She prefers a more continuous, surprising progression rather than a single decisive turning point where humans are abruptly displaced.
AI companions and care: the ‘Geriatric Robot’ story, loneliness, and dignity
A discussion of human-level interaction turns practical and personal: McCorduck’s “Geriatric Robot” concept emphasizes listening and companionship rather than only physical care. She balances enthusiasm for helpful AI with a desire for autonomy and quiet, highlighting the complexities of care and human preference.
Existential risk as ‘male gaze,’ plus women in tech and the sandpile effect
McCorduck reframes some AI doom narratives as ego-threat responses—calling it the “male gaze” applied to intelligence and agency. The conversation expands to her work on women in science/tech, arguing progress is insufficient but can shift suddenly via “sandpile” dynamics, as seen in MeToo.
A memoir’s lesson: luck, optimism, and why she wanted AI to succeed
In closing, McCorduck reflects on writing her memoir and recognizing the extraordinary luck of witnessing AI’s birth firsthand. She reveals a personal, feminist motivation: hoping AI would undermine the myth that intelligence is fundamentally male, and ends with cautious optimism about unknown future risks.
