Reasoning models' success makes denying computationalism ever harder

Aaroth · x · 2026-09-14

The author argues that "the laws of computation bind human cognition just as much as silicon" was mere conjecture a decade ago, but the obvious success of AI — especially recent reasoning models — means denying computationalism now requires ignoring mounting in silico evidence. Admits a non-computational secret sauce is still possible, but the burden of proof has shifted.

Original post →

More from AGI Musings

AGI Musings channel →