@atoponce I wonder: you know how virtual assistants are given feminine names and voices (Siri, Alexa)? And you know how there is a persistant false belief that women are somehow worse at math than men?

I have to wonder whether that combination of biases has any influence on the programmers who create these LLMs? I mean on top of all of the other biases and misunderstandings they already have about neuroscience and language? Are they creating their own stereotype of a ditzy secretary?