Podcast Summary: Alison Gopnik on Childhood Learning, AI as a Cultural Technology, and Rethinking Nature vs. Nurture

Podcast: Conversations with Tyler
Host: Tyler Cowen (Mercatus Center at George Mason University)
Guest: Alison Gopnik (Professor of Psychology and Philosophy, UC Berkeley)
Date: December 17, 2025

Overview

In this episode, Tyler Cowen interviews developmental psychologist and philosopher Alison Gopnik. They explore how children learn and experiment, the parallels between scientific thinking and child cognition, the limitations of nature vs. nurture debates, and what generative AI reveals (and obscures) about intelligence. The discussion also touches on consciousness in children, the validity of traditional models like Piaget and Freud, implications for education, the pitfalls of psychological diagnoses, and the societal undervaluing of caregiving.

Key Discussion Points & Insights

1. Children as Natural Scientists and Bayesian Thinkers

(02:34–11:38)

• Children & Scientists Learn Similarly:
  Gopnik draws parallels between the learning process of children and scientists:

  "Little kids are looking at data and systematically figuring out what kind of structure out there in the world could have caused this pattern of data." (B, 03:46)

• Bayesian Approaches in Children:
  Children often behave in more "Bayesian" ways than adult scientists, especially in adjusting their beliefs based on surprising evidence:

  "Kids actually are better at solving problems that involve unusual outcomes than the scientists are." (B, 05:37)

• Simulated Annealing as a Learning Model:
  Gopnik uses the concept of "simulated annealing" from computer science to describe child cognition:

  "Four-year-olds seem to be a really good idea of this kind of random search." (B, 08:15)

• Children’s Explorations vs. Outcome-driven Behavior:
  Children’s exploratory behaviors, like experimenting with a spoon and avocado, reflect scientific methods more than routine actions:

  "You see them doing these kind of experiments all the time… just try all these different things." (B, 10:14)

2. Theories of Learning: Beyond Minimizing Surprise

(11:38–13:41)

• Critique of ‘Minimize Surprise’ (Friston) Theory:
  Gopnik finds Friston’s “minimize surprise” model attractive but overly broad and empirically slippery:

  "There's this kind of category of theories...that have a good combination of intuitively being in the right direction and then have a lot of math. But it's actually very hard to connect them to experiments." (B, 12:06)

• Emphasis on Intervention and Experimentation:
  She argues that active intervention, not just responding to surprise, is core to genuine learning:

  "It's not just that you're seeing something surprising and then changing your view, it's that you're seeing something surprising. You're doing an intervention... And then that's really the thing that's driving your theory change." (B, 13:10)

3. Optimizing Experimental Studies of Childhood Learning

(13:41–15:43)

• Big Data Approaches with Children:
  If resources were unlimited, Gopnik would use technology (e.g., GoPro cameras) to capture and analyze vast, real-life streams of children’s experiences to discern systematic patterns in their exploratory actions.

• Hypothesis:

  "...much more systematic relationships between what happens, what the baby does next, what the outcome is than you might think on the surface." (B, 15:16)

4. Consciousness in Babies vs. Adults

(15:43–21:23)

• Consciousness: Not a Single Thing:
  Gopnik believes consciousness is multifaceted, akin to "life." Babies are more conscious in terms of richness and receptivity to experience:

  "Babies...are conscious of all the things that are going on around them." (B, 16:36)

• Memory and Consciousness:
  Lesser episodic memory may mean experiencing more in the present:

  "...before they develop a lot of episodic [memory]...that actually makes them more conscious in the sense that they're more focused on the present and they're experiencing the present more." (B, 18:18)

• Aphantasia and Consciousness:
  Not having mental imagery (aphantasia) does not reduce conscious experience. It points to differences in how top-down visual processes work, not diminished visual experience or consciousness.

  "It just shows that there's a big gap between what we think of...and what's actually going on cognitively." (B, 19:33)

5. Babies’ Responses to Animation & Human Visual Processing

(21:23–23:44)

• Babies’ Attraction to Animation:
  Animation (like Pixar films) resonates with babies due to their precocious ability to read faces and emotions, though animators’ expertise is intuitive, not analytical.

• Vision as "Inverse Graphics":
  Some current theories liken human visual processing to computer graphics—building world models from perceptual input.

6. Legacies of Freud and Piaget

(23:44–27:58)

• Freud's Enduring Point:
  The recognition that even young children make inferences about the social world has held up, but Freud’s influence in psychology is now limited.

• Piaget's Lasting Impact:
  Piaget's constructivist model—children build abstract world models from experience—remains foundational:

  "His observations have held up remarkably well...even though his interpretation...has changed a lot over time." (B, 26:38)

7. Nature vs. Nurture: Complexity Beyond Twin Studies

(27:58–38:41)

• Problems with Simple Nature/Nurture Division:
  Twin studies often oversimplify developmental effects; gene-environment interactions are far more complex.

  "If the effect of nurture is not sort of on the mean but on the variation...you're not going to see it in any straightforward way in a twin step." (B, 31:54)

• Wealth, Genetics, and Variation:
  As environments improve, genetic effects might appear more salient, but so does room for variability.

• Focus on Commonality:
  Gopnik finds the universality of human cognitive capacities more interesting than individual differences:

  "...given how unbelievably smart all of them are to begin with, I just think it's not an interesting question." (B, 35:38)

• "Heritability" in Families:
  Gopnik uses her own family as evidence that nurture amplifies differences and variability rather than simply shifting mean outcomes.

  "The effect of caregiving...is to increase variability, is to increase variation..." (B, 57:06)

8. Education: Inquiry, Apprenticeship, and the Apprenticeship Gap

(38:41–42:53)

• Early Childhood:
  Inquiry-based and play-based learning with warm caregiving is optimal.

• School Age (7+):
  Apprenticeship models, akin to music or sports (practice, feedback, demonstration), are better than info-heavy, test-centric systems:

  "Imagine if we tried to teach baseball the way that we teach science...you kind of wouldn't think that people would be as good at baseball." (B, 40:30)

• Goodhart’s Law in Schooling:
  The system often selects for students who master "doing school", not for creativity or genuine inquiry.

9. Generative AI as Cultural Technology, not Superintelligence

(42:53–51:55)

• AI as Cultural Technology:
  Gopnik argues that generative AI is more like books, libraries, or print than a new form of intelligent agency:

  "It's a way of getting information from other people...the latest technique for doing that." (B, 43:27)

• Skepticism about AI Reasoning:
  She doubts that current reasoning models equate to genuine or novel reasoning:

  "I would be impressed if they were actually designing experiments that would tell you something about something new...that all the other people around them didn't know before." (B, 46:19)

• Limitations in Physical and Experimentation Abilities:
  LLMs excel at combining existing knowledge but lag in tasks requiring novel experimentation or real-world intervention—areas where humans, even toddlers, excel.

• Hallucination & Objective Functions:
  AI’s “hallucinations” occur because their goal is to generate acceptably coherent and pleasing text, not truth.

• AI Outperforming Humans:
  Cowen challenges Gopnik with cases where AI exceeds humans on specific tasks, including economic problems, but Gopnik maintains skepticism about genuine creativity and novel insight.

10. Rethinking Psychological Diagnostics: Autism & ADHD

(51:55–55:30)

• Autism:
  Gopnik sees "autism" not as a singular condition but a broad range of variations, analogous to obsolete diagnoses like "dropsy":

  "It's just too simple. Again, to use another analogy, it's as if in the 19th century you said, do people with dropsy have some characteristic or not? And it turns out dropsy is not actually a thing, dropsy is a symptom..." (B, 52:52)

• ADHD & Societal Context:
  Variation in attention is natural, but industrial/school societies valorize focused attention, making broader attentiveness seem dysfunctional in context.

11. Society and Caregiving: The Overlooked Core

(59:10–60:41)

• Caregiving’s Invisible Value:
  Gopnik emphasizes that caregiving (for children, elders) is central to meaning in life but undervalued by both economics and social sciences:

  "All that work that we do, taking care of each other, it's just completely invisible from an economic perspective." (B, 60:16)

• Current/Future Work:
  Gopnik’s forthcoming work and book will focus on the role of caregiving across the lifespan, especially relating to elders.

Notable Quotes and Memorable Moments

• On Children’s Learning:
  "Four-year-olds seem to be a really good idea of this kind of random search." – Alison Gopnik (08:15)

• On "Nature vs. Nurture":
  "If the effect of nurture is not sort of on the mean but on the variation...you're not going to see it in any straightforward way..." – Alison Gopnik (31:54)

• On the Impact of Caregiving:
  "What nurture will do is let you have variability. That's the thing that is, in a sense, is heritable." – Alison Gopnik (58:08)

• On AI as Cultural Technology:
  "What generative AI tells you is, here's a summary of what all the people on the net have said...learning how to use those cultural technologies..." – Alison Gopnik (43:27)

• On Educational Methods:
  "Imagine if we tried to teach baseball the way that we teach science...you kind of wouldn't think that people would be as good at baseball." – Alison Gopnik (40:30)

• On Piaget’s Enduring Relevance:
  "His observations have held up remarkably well...even though his interpretation...has changed a lot over time." – Alison Gopnik (26:38)

Timestamps for Key Segments

| Timestamp | Topic | |-----------|-------| | 02:34–11:38 | Children as scientists, Bayesian reasoning, simulated annealing metaphor | | 13:41–15:43 | Designing big data experiments in developmental psych | | 15:43–21:23 | Consciousness in babies, episodic memory, aphantasia | | 21:23–23:44 | Animators, vision theories, and children's perception | | 23:44–27:58 | Freud and Piaget's relevance today | | 27:58–38:41 | Nature vs. nurture debate; complexity of gene-environment interplay | | 38:41–42:53 | Improving education: early childhood, apprenticeship models, critique of schooling | | 42:53–51:55 | Generative AI’s nature, limits, hallucination, and comparison to print/internet | | 51:55–55:30 | Autism, ADHD, and diagnostic ambiguity | | 59:10–60:41 | The centrality and societal neglect of caregiving |

Tone and Closing Thoughts

The conversation is intellectually rich, skeptical of dogma, and frequently playful—true to Tyler Cowen’s probing style and Gopnik’s reflective, theory-challenging approach. Gopnik insists that the complexity of development, learning, intelligence, and social relationships elude simple frameworks and dichotomies.

For Further Exploration:

• Gopnik’s books: The Gardener and the Carpenter, forthcoming work on caregiving
• Recent work with Henry Farrell and James Evans on AI as cultural technology
• Piaget’s original developmental studies
• Eric Turkheimer's research on genetics and socioeconomic status