Grady Booch says AI–brain “convergent evolution” is only an amusing parallel

Grady_Booch · x · 2026-07-23

Grady Booch criticizes a paper claiming “convergent evolution” between AI models and brains, arguing that the two systems evolved under entirely different reward functions and external pressures. He says the result is, at best, an amusing parallel—not a meaningful equivalence.

He is replying to a Substack post that frames the paper as a mathematical theory of convergence via contravariance and “zippering theorems.”

Related event: NeuroAI Paper Sparks Debate: Is AI-Brain Convergent Evolution Inevitable or Just an Analogy?(7 posts)→

Original post →

More from AGI Musings

AGI Musings channel →