Article by The Verge and New York Magazine, documenting the hidden work that goes into “training” AI and the conditions of the people who do that.
You must log in or # to comment.
You get the feeling that the ‘we have to choose between two shitty answers’ problem mentioned is because the chatbots being trained are trying to be everything at once?
You are probably on to something there. Chasing the dream of a general purpose chatbot seems to be what the venture capitalists demand though, so we will be stuck with them…
