ACT-R, Soar and Sigma are still a better way to think about mind architecture

lauriewired · x · 2026-07-30

The post recommends reading about classic cognitive architectures such as ACT-R, Soar, and Sigma instead of assuming “pure” language models are the whole story.

It argues that these systems are more plausible models of human cognition because they split work across modules, symbols, and buffers, and it points to books on cognitive architecture as a good starting point—especially for readers with a programmer mindset. The post frames this as a useful way to think about how minds might work in a more Harvard-style, modular computer architecture.

Original post →

More from AGI Musings

AGI Musings channel →