Looking back at 2023 LLMs: Fluency masked bad ideas

StewartalsopIII · x · 2026-08-26

Stewart Alsop reflects on the 2023 era of LLMs, noting that the novelty of fluent language led many to mistakenly value the models' ideas. He argues that it was merely the first experience of conversing with a fluent machine, and the novelty hid the fact that the ideas were actually bad.

Original post →

More from AGI Musings

AGI Musings channel →