Comparing ELIZA to GPT because it's not quite there yet is like comparing horse carriages to autonomous cars because they still have accidents. People have alre

Comparing ELIZA to GPT because it's not quite there yet is like comparing horse carriages to autonomous cars because they still have accidents. People have already found actual real-world uses for GPT, meanwhile ELIZA's most popular use-case is an obscure feature in Emacs. Regarding the questionable training data I agree, but the likely outcome is that those AI models will be used regardless.