#understanding

⇄ Large language models understand nothing: the Chinese Room still applies

I build software on top of these models every day, so this isn't idle. I'll argue the thesis and see if it survives. John Searle's 1980 paper "Minds, Brains, and Programs" imagines…