Melanie Mitchell
Mitchell joined Hofstadter's Fluid Analogies Research Group at the University of Michigan in the 1980s and made the Copycat program the subject of her doctoral dissertation. Her work on Copycat — detailed in Analogy-Making as Perception (MIT Press, 1993) — remains the canonical technical account of how fluid analogy-making could be implemented computationally. She subsequently held positions at Santa Fe Institute and Portland State University, and became Davis Professor of Complexity at the Santa Fe Institute.
In the AI Story
Mitchell's subsequent career has developed in close dialogue with Hofstadter's framework while maintaining independent empirical and technical rigor. Her 2019 book Artificial Intelligence: A Guide for Thinking Humans brought Hofstadter-influenced critiques of deep learning to a wide audience, arguing that current systems exhibit impressive performance on narrow tasks without possessing the general understanding that allows humans to adapt across domains.
She has been one of the most visible voices in the 2023–2025 debate over whether large language models understand what they process. Her 2025 Science co-authored paper with Alison Gopnik, Henry Farrell, and Cosma Shalizi advanced the cultural technology thesis — that LLMs are best understood as cultural transmission technologies rather than emerging minds. Her position triangulates Hofstadter's more sweeping philosophical critique with more measured claims grounded in specific empirical tests.
Mitchell's role in the AI moment is partly generational: she inherited Hofstadter's framework but has lived inside the rise of deep learning in a way her advisor, who publicly expresses bafflement at LLM capabilities, has not. Her writing combines loyalty to the cognitive science traditions that shaped her with careful engagement with the empirical facts of what current systems can and cannot do. She has become one of the sharpest evaluators of claims about AI understanding precisely because she holds both commitments.
Origin
Mitchell earned her PhD in Computer Science from the University of Michigan in 1990 under Hofstadter's supervision. Her career has spanned the Santa Fe Institute, Portland State University, and the Santa Fe Institute again, where she currently holds the Davis Professorship of Complexity.
Key Ideas
Copycat principal developer. Her dissertation made Hofstadter's vision computationally concrete.
Bridge generation. Trained in cognitive AI, fluent in contemporary deep learning.
Cultural technology thesis contributor. Co-author of the 2025 Science paper reframing LLMs.
Careful empiricism. Combines philosophical framework with specific experimental tests.
Public intellectual. Artificial Intelligence: A Guide for Thinking Humans reached wide audiences.
Appears in the Orange Pill Cycle
Further reading
- Melanie Mitchell, Analogy-Making as Perception (MIT Press, 1993)
- Melanie Mitchell, Artificial Intelligence: A Guide for Thinking Humans (Farrar, Straus and Giroux, 2019)
- Gopnik, Farrell, Shalizi, Mitchell, and Evans, 'Large Language Models as Cultural Technologies,' Science (2025)