The assertion that LLMs are "capable of surprisingly sophisticated reasoning" is supported with a link an article @willknight wrote on the "Sparks of AGI" paper + criticism of it. >>

→ View original post on X — @emilymbender