Lex Fridman PodcastPamela McCorduck: Machines Who Think and the Early Days of AI | Lex Fridman Podcast #34
CHAPTERS
- 0:00 – 3:25
Why McCorduck wrote "Machines Who Think": from novelist to AI historian
Lex introduces Pamela McCorduck and the premise of her 1979 book, which treats AI as both a technical field and a mythic human impulse. McCorduck describes how a would-be AI novel became a far more demanding project: a living history built from interviews with the field’s founders.
- •Started as a novelist; initially imagined writing an AI-themed novel
- •Realized no one was documenting AI’s early history; chose interviews and narrative history
- •Early skepticism from John McCarthy and others about the value of the project
- •The challenge of writing history while the field was still forming
- 3:25 – 4:17
Meeting the founders while the field was still young
McCorduck explains why it was uniquely valuable to write AI history while the pioneers were still active researchers. She reflects on their confidence and ambition, and on the generosity of their time despite busy careers.
- •Pitch: the founders were ‘alive and kicking’ and could speak for themselves
- •Founders understood they were doing something important
- •Interviews happened while many were at the height of their research careers
- •The unusual dynamic of a writer documenting an emerging scientific community
- 4:17 – 5:18
Dartmouth legacy and the four ‘founding fathers’ of AI
Asked to “paint a picture” of the era, McCorduck identifies the core figures who defined the field’s early direction. She highlights Newell and Simon’s Logic Theorist as a decisive proof-of-concept that set them apart at Dartmouth.
- •Founding fathers: Newell, Simon, Minsky, McCarthy
- •Newell & Simon arrived with a working program: Logic Theorist
- •Others had ideas, but not working systems yet
- •Different motivations and styles among the founders
- 5:18 – 11:22
Her path into AI: Stanford, Feigenbaum, and early AI literature
McCorduck recounts her immersion in AI through Stanford and Ed Feigenbaum, including work on one of the first AI reader textbooks. As an English major, she was captivated by the possibility of intelligence outside the human brain, which shaped her literary-historical framing.
- •Worked as Ed Feigenbaum’s assistant at Stanford
- •Helped on 'Computers and Thought' (early AI readings/textbook)
- •Fascination with ‘intelligence outside the human cranium’
- •Her humanities background helped connect AI to long cultural continuities
- 11:22 – 12:59
Mythic roots of AI: Homer, the Golem, Babbage, and a long human obsession
The conversation widens beyond 20th-century computing into ancient and pre-digital antecedents. McCorduck argues that the desire for artificial beings predates modern science, appearing in classical epics and later in early mechanical and computing visions.
- •Robots and automation themes in Homer’s Iliad and Odyssey
- •Legendary roots: the Golem and early 20th-century ‘robots’
- •19th/early 20th century attempts lacked enabling technology
- •Babbage and Lovelace anticipating broader capabilities of computation
- 12:59 – 14:22
Hellenic vs Hebraic attitudes: welcoming robots vs fearing blasphemy
McCorduck contrasts two cultural lenses for interpreting artificial beings: one celebratory and utilitarian, the other prohibitive and moralizing. She links enduring public anxiety about AI to deep-seated religious and cultural narratives, while rejecting ‘blasphemy’ as the real issue.
- •Hellenic view: robots as helpful assistants (e.g., Hephaestus’ helpers)
- •Hebraic view: prohibition on imitation (‘graven images’) and fear of overstepping
- •AI anxiety often framed as wicked or blasphemous
- •Downsides exist, but not because it is ‘forbidden’ in a religious sense
- 14:22 – 23:35
Frankenstein, machine villains, and the deeper psychology of AI fear
Lex and McCorduck explore why AI fears feel primal, beyond practical concerns like bias or algorithmic manipulation. McCorduck and McCarthy’s ‘literary problem’ frames a recurring narrative: humans as heroes, machines as villains—while Frankenstein complicates who the true monster is.
- •Modern fears: bias, manipulation, and social control vs deeper existential dread
- •Primal fear: machines will outthink and replace humans
- •McCarthy’s ‘literary problem’: conventions cast machines as villains
- •Frankenstein: a being seeking love becomes monstrous through rejection; parallels to dehumanization and abolition discourse
- 23:35 – 28:46
AI as an ‘outcast’ and the CMU social world: Knuth, Traub, and Herb Simon’s sherry salon
McCorduck describes moments when AI was marginalized within computer science and needed outside validation. She then shifts to the personal community around CMU—especially her friendship with Herb Simon—and the tradition of deep conversation and salons bridging science and the humanities.
- •1970s institutional skepticism: proposals to exclude AI from defining computer science
- •Don Knuth’s intervention: insisted AI belonged in the story of computer science
- •Joe Traub (algorithmic complexity) praised AI researchers as exceptionally smart
- •Friendship with Herb Simon; conversations on literature, music, art, and life
- •Creation of a monthly discussion group with Newell/Simon and writers in Pittsburgh
- 28:46 – 37:08
Symbolic AI, deep learning, and why she rejects the ‘AI winter’ narrative
McCorduck reflects on how AI’s center of gravity shifted from symbolic systems to algorithmic approaches, surprising her. She argues “AI winter” is largely a commerce-driven story about monetization and hype cycles, not a halt in basic research—though funding shocks (e.g., Britain’s Lighthill Report) can cause real damage.
- •Her early skepticism about whether AI would ‘work out,’ and how progress surprised her
- •Symbolic AI ‘dried up’ while algorithmic methods became the main event
- •Bias and human bigotries can be ‘baked into’ algorithmic systems
- •AI winter as a ‘crock’: basic research remained important; commercialization lagged
- •Funding politics can cripple science (Lighthill Report); later resurgence with DeepMind
- 37:08 – 43:32
Santa Fe Institute and complexity: finding new language for creative AI (Harold Cohen’s AARON)
McCorduck connects AI to complexity science through her experience at the Santa Fe Institute. Work on Harold Cohen’s AARON led her to see artistic cognition as a complex adaptive system, and Santa Fe gave her the vocabulary and community to make sense of those connections.
- •Harold Cohen’s AARON as a reflection of cognitive process and creativity
- •Personal friction and the feeling of being ‘finished’ with AI writing
- •Santa Fe Institute experience (early 1990s) as intellectually restorative
- •Open-door culture: learning from Kauffman, Gell-Mann, and others
- •Regret at not writing a book about SFI when invited by George Cowan
- 43:32 – 54:08
Modern AI excitement, the singularity debate, care robots, and the ‘male gaze’ of existential risk
The discussion turns to today’s AI boom, timelines, and public narratives. McCorduck dismisses Kurzweil’s singularity framing as a dramatic ‘game over’ moment, proposes more human-centered visions like listening care robots, and critiques existential-risk rhetoric as partly driven by threatened ego—while emphasizing ethics, empathy, and unknown unknowns.
- •Today’s AI community feels like earlier eras: excitement and belief in impact
- •Breakthrough timelines are unpredictable (e.g., Hinton’s 1986 ideas needing later compute/data)
- •Skepticism about Kurzweil-style singularity as a single decisive moment
- •The ‘Geriatric Robot’: value in listening and companionship, not just tasks
- •Existential risk discourse framed as ‘the male gaze’—fear of being surpassed
- •Key worry: unforeseen, unanticipated consequences; need broader notions of intelligence including ethics and empathy
- 54:08 – 1:00:06
Women in tech, MeToo as a ‘sandpile’ tipping point, and her personal motive for wanting AI to succeed
McCorduck discusses persistent barriers for women in computing and revisits her book ‘The Futures of Women,’ emphasizing slow progress and backlash dynamics. She ends with a striking personal insight: she subconsciously rooted for AI because it challenged the myth that intelligence ‘resides in the male cranium.’
- •Progress for women in science/tech has been insufficient; systemic grinding down persists
- •Four ‘futures’ as facets of the present: backlash and incremental gains
- •‘Golden age of equality’ and the sandpile effect; MeToo as a tipping event
- •Congenital optimism despite slow change
- •Personal revelation: AI as a counterexample to gendered notions of intelligence
