LLMs reveal the latent shadow of structure in language

kalomaze · x · 2026-08-16

The author posits that language models have demonstrated the latent structure within our speech is a sufficient raw material for continued progress. The discussion explores taking this claim seriously and searching for valuable, broken 'invisible assumptions'.

Related event: AI Breakthroughs Stem from Breaking Implicit Assumptions(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →