"Don't confuse mimicry with intelligence": the LLM smartness debate resurfaces

AlexTensor · x · 2026-11-19

A widely shared take argues LLMs are as "smart" as smartphones—smart in name only—and that mimicry should not be confused with intelligence; LLMs plus harnesses can be useful, but intelligent they are not. The quote references a longer thread's framing: one camp (AI doomers) treats intelligence as one-dimensional, so improving LLMs means approaching AGI; another treats it as binary, so current progress says nothing about general intelligence. The debate centers on how to interpret the dimensionality of intelligence and what LLM progress actually means.

Original post →

More from AGI Musings

AGI Musings channel →