for now, chatbots tend to sycophantically confabulate upon and reinforce user prejudices and inclinations, rather than reining them in. there’ve been prominent cases of people going a bit mad this way. 1/

that could be remedied. chatbots could have some version of consensus reality towards which they guide users. but then these massive, centralized, profit-seeking companies that run the chatbots would largely define that “consensus”. 2/

do you think “touching grass” with that version of reality would, eg, dissuade people from voting for Donald Trump, or, less politically, force them to confront realities like climate change is real and vaccines work and tylenol has not been shown to cause autism? /fin